Skip to main content

Full text of "Respiratory care : the official journal of the American Association for Respiratory Therapy"

See other formats

June 1993 
Volume 38, Number 6 

ISSN 009891 42-RECACP 




Oxygenation in the Critically III Patient 


Normal and Abnormal Oxygenation 

Clinical Effects of Hypoxemia and 
Tissue Hypoxia 

When Does Vo 2 Depend on Do 2 ? 

Oxygenation Indices 

Techniques and Devices for Monitoring 

Devices and Techniques for Increasing 
Arterial Oxygenation 

Aggressive Therapy for Oxygenation 


39th Annual Convention and Exhibition 
December 1 1-14, 1993 • Nashville, Tennessee 


Clinically Speaking 

66 In Pressure 
Regulated Volume 
Control, as the 
patient's compliance 
changes, the Servo 
Ventilator 300 with its 
real-time control, adjusts 
to find the lowest 
possible pressure to 
deliver the guaranteed 
volume. 99 

Attending Physician, SICU 

66 Recently, a patient 
who had ARDS was 
placed on the ventilator 
in the PRVC mode. 
We were able to 
ventilate her with 
consistently low peak 
airway pressures, 
and she improved at 
a more rapid pace than 
we anticipated. §9 

66 The flexibility of the 
Servo 300 is superior. 
With a simple turn of a 
dial, you can make 
ventilatory adjustments 
and move between 
different modes. 95 

66 Just as the 
Servo 900C did when 
it was introduced, the 
gas delivery system of 
the Servo 300 has set 
a new standard. 99 

Wilmington, North Carol! 

66 In volume 
support, the Servo 
300 automatically 
fine-tunes the pressure 
support level breath- 
by-breath, and frees 
the therapist for 
other duties. It's 
much more 
efficient. §5 

66 Patients feel 
more comfortable on 
the Servo Ventilator 
300. And it can be 
used with neonates, 
pediatrics, and adults 
one ventilator, 
versus many 
ventilators. §5 

Attending Physician, SICU 
University ol 

Siemens can also bring you 
the very best in ventilator care. 
For more information on 
the Servo Ventilator 300® 
or to arrange a personal 
demonstration, contact your 
local Siemens representative. 
Or call toll-free. 

Siemens Medical Systems, Inc. 

Patient Care Systems Division 
16 Electronics Avenue 
Danvers, MA 01923 
Toll-Free 1-800-333-8646 


technology in caring hands. 

Circle 114 on reader service card 

Put Human Resources 
to Work For You 

The AARC Human Resources Survey: 
A Study of Respiratory Care Human Resources in Hospitals 

Covers a wide range of human resource issues, including compensation, numbers of full-time equivalents, job vacancy 
rates, education, credentialing, and licensure. Even includes information on age, sex, and years of experience. 

• Comprehensive Summary 

• Position Profiles 

• Salaries 

• Education, Experience, and Credentials 

• Regional Demographics 

• Vacancies 

Fall 1992, 68 pages, 66 tables 

Item BK12 $50 Ea (AARC Member $35) 

Also Available 

A Study of Chronic Ventilator-Dependent Patients 

Chronic ventilator-dependent patients are costing American hospitals more than $9 million per day according to this 
Gallup study conducted for the AARC. This important study provides information on patients who depend on life- 
support systems, why, how, and where they are being treated, and the cost of treatment. 

47 pages, 9 tables, 12 figures Item BK20 $50 Ea (AARC Member $25) 

To Order Call (214) 243-2272 or Fax to (214) 484-2720 
Purchase Order • MasterCard • Visa 

Please send me the following: 

□ Human Resources Survey- Item BK1 2— $50 (Member $35) 

□ Chronic Ventilator Patient Survey - Item BK20 — $50 (Member $25) 

Add $3 for Shipping and Handling 

Charge to my Q Visa □ MasterCard Card expires 

Card # Signature 

Payment enclosed $ Charge to Purchase Order # 

AARC Member No. 




American Association for Respiratory Care • 11030 Abies Lane • Dallas, Texas 75229-4593 

What if you 


aerosol therapy? 


What if you were going to use an aerosol compressor every day for the next 

five years. Or longer. If you were going to look at it. Listen to it. If you were going 

to keep it in your home. Carry it around. Take it along when you go out. If you were 

going to rely on it. You'd choose the one that's sleek and contemporary looking. 

Unobtrusive. Significantly more compact than anything else available. The one 

that weighs only 5.5 lbs. 

If you needed aerosol therapy, you'd choose the Passport™ aerosol compressor 
for the way it looks and travels. And then, over time, you'd come to appreciate 
little things like the re-usable inlet filter. And you'd feel very good about the 
5 year compressor warranty. Look at Passport through a patient's eyes. Then 
give your customers what you'd want for yourself. Give them Passport. 
For more information, contact your Invacare sales representative or call 1-800-333-6900. 

r /po/"»fc 

Invacare Corporation 

899 Cleveland Street 

Elyria, Ohio 44036 

©1993 Invacare Corporatic 

Circle 123 on reader service card 


A Monthly Science Journal. Established 1956. Official Journal of the American Association tor Respiratory Care 


11030 Abies Lane 
Dallas TX 75229 

f^OlVTTTlVT^ June 1993 
\^\Jll JrilllkJ Volume 38, Number 6 



Pal Brougher RRT 


Philip Kittredge RRT 



Donna Stephens BBA 



Dean Hess MEd RRT. Chairman 


Thomas A Barnes EdD RRT 

Richard D Branson RRT 


Robert L Chatburn RRT 

Charles G Durbin Jr MD 

Thomas D East PhD 

Robert M Kacmarek PhD RRT 

The Proceedings of a Conference 

Neil R Maclntyre MD 
David J Pierson MD 

held October 8-10, 1992, 

James K Stoller MD 

in Puerto Vallarta. Mexico 


Dean Hess MEd RRT and David J Pierson MD 

Frank E Biondo BS RRT 

Howard J Birenbaum MD 

Chairmen and Guest Editors 

John G Burford MD 

Bob Demers BS RRT 

Donald R Elton MD 

Robert R Fluck Jr MS RRT 

Ronald B George MD 

James M Hurst MD 

Charles G Irvin PhD 


MS Jastremski MD 

Hugh S Mathewson MD 


Normal and Abnormal Oxygenation: Physiology and Clinical 

Michael McPeck BS RRT 


Richard R Richard BS RRT 

John Shigeoka MD 
R Brian Smith MD 

by David J Pierson — Seattle, Washington 

Jack Wanger MBA RPFT RRT 
Jeffrey J Ward MEd RRT 


Clinical Effects of Hypoxemia and Tissue Hypoxia 


by Thomas L Higgins and Jean-Pierre Yared — Cleveland, Ohio 

Stephen M Ayres MD 
Reuben M Cherniack MD 


When Does Vo 2 Depend on D02? 

Joseph M Civetta MD 

by P Terry Phang and James A Russell — Vancouver, British Columbia, 

John B Downs MD 
Donald F Egan MD 


Gareth B Gish MS RRT 

George Gregory MD 


Assessment of Oxygenation: Oxygenation Indices 

Ake Grenvik MD 

H Frederick Helmholz Jr MD 

by Loren D Nelson — Nashville, Tennessee 

John E Hodgkin MD 
William F Miller MD 


Techniques and Devices for Monitoring Oxygenation 

Elian J Nelson RN RRT 
Thomas L Petty MD 

by Dean Hess and Robert M Kacmarek — Boston, Massachusetts 

Alan K Pierce MD 
Henning Pontoppidan MD 


The Nuts and Bolts of Increasing Arterial Oxygenation: Devices and 

John W Severinghaus MD 
Barry A Shapiro MD 


by Richard D Branson — Cincinnati, Ohio 


Linda Barcus 
Sieve Bowden 


The Magic Bullets in the War on ARDS: Aggressive Therapy for 

Bill Cryer 

Oxygenation Failure 

Donna Knauf 
Jeannie Marchant 

by Thomas D East — Salt Lake City, Utah 

Respiratory Care (ISSN 00989142) is a monthly publication of Daedalus Enterprises Inc for the American Association lor Respiratory Care. Copyright ° 1993 by Daedalus En- 
terprises Inc, 1 1030 Abies Lane. Dallas TX 75229. All rights reserved. Reproduction in whole or in part without the express, written permission of Daedalus Enterprises Inc, is 
prohibited. The opinions expressed in any article or editorial are those of the author and do not necessarily reflect the views of Daedalus Enterprises Inc. the Editorial Board, or 
the American Association for Respiratory Care. Neither can Daedalus Enterprises Inc, the Editorial Board, or [he Amercian Association for Respiratory Care be responsible for 
the consequences of the clinical applications of any methods or devices described herein. Printed in USA. 
Respiratory Care is indexed in Hospital Literature Index and in Cumulative Index to Nursing and Allied Health Literature. 
Subscription Rales: $5.00 per copy; $50.00 per year (12 issues) in the US: $70.00 in all other countries (add $84.00 for airmail). 
Second Class Postage paid at Dallas, TX. POSTMASTER: Send address changes to Respiratory Care. Daedalus Enterprises, Inc., 11030 Abies Lane. Dallas TX 7522". 




a reason 

why more than half 

the pulse oximeters 

in the world 

are Nellcors. 


For information about Nellcor's complete family of 
pulse oximeters and sensors, contact your local Nellcor 
representative or call 1-800-NELLCOR. Nellcor Incorporated, 
25495 Whitesell Street, Hayward, CA 94545, USA, telephone 
510 887-5858. 

Internationally, contact our European office in the 
Netherlands at +31. 73.426565 or the Asia/Pacific office in 
Hong Kong at +852.529.0363. 


Circle number 113on Reader Service Card 


Instructions for Authors and Typists is printed 
near the end of Respiratori Care on a quarterly 
basis (Jan, Apr, July, Nov). 


PHOTOCOPYING. Any material in this journal 
that is copyrighted by Daedalus Enterprises. Inc 
may be photocopied for noncommericial pur- 
poses of scientific or educational advancement. 

QUOTATION. Anyone may, without permis- 
sion, quote up to 500 words of material in this 
journal that is copyrighted by Daedalus En- 
terprises Inc, provided the quotation is for non- 
commercial use. and provided Respiratory Care 
is credited. Longer quotation requires written ap- 
proval by the author and publisher. 


Respiratory Care 
11030 Abies Lane 
Dallas TX 75229-4593 


SUBSCRIPTIONS. Individual subscription rates 
are $50.00 per year (12 issues) in the U.S. and 
Puerto Rico. $70.00 per year in all other coun- 
tries; $95.00 for 2 years in the U.S. and Puerto 
Rico. $135.00 in all other countries; and $140 
for 3 years in the U.S. and Puerto Rico. $200.00 
in all other countries (add $84.00 per year for air 
mail). Annual organizational subscriptions are 
offered to members of associations according to 
their membership enrollment as follows: 101- 
500 members— $5.00, 501-1.500 members— 
$4.50, 1,501-2,500 members— $4.25. 2.501- 
5,000 members— $4.00, 5,001-10,000 mem- 
bers— $3.00, and over 10,000 members— $2.50. 
Single copies, when available, cost $5.00; add 
$7.00 air mail postage to overseas countries. 

CHANGE OF ADDRESS. Six weeks notice is 
required to effect a change of address. Note your 
subscription number (from the mailing label) 
your name, and both old and new address, in- 
cluding zip codes. Please note your subscription 
number on the envelope. Copies will not be re- 
placed without charge unless request is received 
within 60 days of the mailing in the U.S. or with- 
in 90 days in other countries. 


Dale Griffiths 

Beth Binkley 

ADVERTISING. Display advertising should be 
arranged with the advertising representatives. 
Respiratory Care does not publish a classified 
advertising column. 


Aries Advertising Representatives 
4 Orchard Hill Road 
Marlboro NJ 07746 
fax (908) 946-1229 


Valley Forge Press 
1288 Valley Forge Rd. Suite 50 
Valley Forge PA 19481 
(800)'220-4979 • (215) 935-3301 
fax (215) 935-8208 



June 1993 
Volume 38, Number 6 


562 Summaries of Pertinent Articles from Other Journals 


707 Exercise Testing System 

707 Breathlessness Book 

707 Home Respiratory Care Poster 

707 Bedside pH and Blood Gas Analyzer 


706 Meeting Dates, Locations, Themes 


705 Examination Dates, Notices, Prizes 


708 Authors in This Issue 
708 Advertisers in This Issue 
564 Advertiser Help Lines 




Summaries of Pertinent Articles in Other Journals 

Editorials, Commentaries, and Special Articles To Note 

Timing of Tracheotomy in Mechanically Ventilated Patients (clinical 
commentary)— JE Heffner. Am Rev Respir Dis 1993;147:768-771. 

Cystic Fibrosis Foundation Consensus Conference Report on Pul- 
monary Complications of Cystic Fibrosis — DV Schidlow, LM Taus- 
sig, MR Knowles. Pediatr Pulmonol 1993;15:187-198. 

International Standards for Safety in the Intensive Care Unit (spe- 
cial article) — The International Task Force on Safety in the Intensive 
Care Unit. Crit Care Med 1993;21:453-456. 

Inhaled Nitric Oxide: The Past, The Present, and the Future (editori- 
al)— RG Pearl. Anesthesiology 1993;78:413-416. (Pertains to Frostell et 
al paper abstracted on Page 568.) 

Pulse Oximetry as a Standard of Practice in Anesthesia (editorial) — 
JH Eichhorn. Anesthesiology 1993;78:423-426. (Pertains to Moller et al 
papers abstracted on Pages 568 and 571 ). 

The Sick Building Syndrome in Office Buildings: A Breath of Fresh 

Air (editorial)— K Kreiss. N Engl J Med 1993;328:877-878. (Pertains to 
Menzies et al paper abstracted on Page 571.) 

Sleep Apnea: A Major Public Health Problem (editorial) — EA Phillip- 
son. N Engl J Med 1993;328:1272-1273. (Pertains to Young et al paper 
abstracted on Page 572.) 

So Who Needs a Bronchoscope, Anyway? (editorial) — RE Wood. Pe- 
diatr Pulmonol 1993; 15:204. (Pertains to Koumbourlis and Kurland 
paper abstracted on Page 572.) 

Nasal Continuous Positive Airway 
Pressure in Chronic Heart Failure 
with Sleep-Disordered Breathing 

— RJO Davies, KJ Harrington, OJM 
Ormerod, JR Stradling. Am Rev Re- 
spir Dis 1993:147:630. 

Nasal continuous positive airway 
pressure (NCPAP) has been reported 
to improve daytime symptoms in pa- 
tients with sleep disordered breathing 
due to heart failure. To examine this 
in a controlled manner, eight men 
with stable chronic heart failure 
(mean left ventricular ejection frac- 
tion 18% and mean frusemide dose 
160 mg) were entered into a con- 

trolled trial of domiciliary nocturnal 
NCPAP. At polysomnography (with 
sleep apnea quantified as the number 
of > 4% dips in arterial saturation per 
hour), seven had nocturnal Cheyne- 
Stokes respiration (S a o : dip rate 3 to 
27/hr), and one both central and ob- 
structive apneas (S a o ; dip rate 8/hr). 
After 2 wk nocturnal domiciliary 
NCPAP at < 1.5 cm H 2 (placebo) 
and 7.5 cm H:0 (active) in random 
order, bicycle exercise tolerance and 
heart failure symptoms (modified Li- 
ken questionnaire) were assessed by 
an observer unaware of the patients' 
NCPAP status. Pulse oximetry (all 
subjects) and radionuclide estimated 

left ventricular ejection fraction 
(three subjects) were also measured 
at the end of each period. Two sub- 
jects withdrew from the study be- 
cause of worsening heart failure dur- 
ing active NCPAP (7.5 cm H 2 0). and 
one of these subjects died. In the re- 
maining six subjects exercise toler- 
ance, symptom scores, and the sever- 
ity of sleep apnea were similar on 
active NCPAP compared with pla- 
cebo. When it was measured, resting 
left ventricular ejection fraction was 
lower on active therapy than on pla- 
cebo. These data exclude a 25% im- 
provement in exercise tolerance with 
95% confidence and suggest that a 




intratracheal suspension 

bovine pulmonary surfactant 


Sterile Suspension 
For Intratracheal 
Administration Only - 



From Ross Laboratories — 
Helping Premature Babies Survive 

Please see adjacent column for Brief Summary of prescribing information. 

Circle 125 on reader service card 

SURVANTA® (1040) 


intratracheal suspension 

Sterile Suspension/For Intratracheal Use Only 


SURVANTA is indicated tor prevention and 
treatment ("rescue") ot Respiratory Distress 
Syndrome (RDS) (hyaline membrane disease) 
in premature infants SURVANTA significantly 
reduces the incidence of RDS, mortality due to 
RDS and air leak complications 

In premature infants less than 1250 g birth 
weight or with evidence of surfactant defi- 
ciency, give SURVANTA as soon as possible, 
preferably within 15 minutes of birth. 

To treat infants with RDS confirmed by x-ray 
and requiring mechanical ventilation, give 
SURVANTA as soon as possible, preferably by 
8 hours of age. 
None known 
SURVANTA is intended for intratracheal use only 

fore, its use should be restricted to a highly 
supervised clinical setting with immediate 
availability of clinicians experienced with intu- 
bation, ventilator management, and general 
care of premature infants Infants receiving 
SURVANTA should be frequently monitored 
with arterial or transcutaneous measurement 
of systemic oxygen and carbon dioxide. 

HAVE BEEN REPORTED If these occur, stop 
the dosing procedure and initiate appropriate 
measures to alleviate the condition After sta- 
bilization, resume the dosing procedure 

Rales and moist breath sounds can occur 
transiently after administration Endotracheal 
suctioning or other remedial action is not 
necessary unless clear-cut signs of airway 
obstruction are present 

Increased probability of post-treatment 
nosocomial sepsis in SURVANTA-treated 
infants was observed in the controlled clinical 
trials (Table 3) The increased risk for sepsis 
among SURVANTA-treated infants was not 
associated with increased mortality among 
these infants. The causative organisms were 
similar in treated and control infants. There 
was no significant difference between groups 
in the rate of post-treatment infections other 
than sepsis. 

Use of SURVANTA in infants less than 600 g 
birth weight or greater than 1750 g birth 
weight has not been evaluated in controlled 
trials There is no controlled experience with 
use of SURVANTA in conjunction with experi- 
mental therapies for RDS (eg, high-frequency 
ventilation or extracorporeal membrane 

No information is available on the effects of 
doses other than 100 mg phospholipids/kg, 
more than four doses, dosing more frequently 
than every 6 hours, or administration after 
48 hours of age. 
Carcinogenesis, Mutagenesis, 
Impairment ol Fertility 
Reproduction studies in animals have not been 
completed Mutagenicity studies were nega- 
tive. Carcinogenicity studies have not been 
performed with SURVANTA 

The most commonly reported adverse experi- 
ences were associated with the dosing pro- 
cedure In the multiple-dose controlled 
clinical trials, transient bradycardia occurred 
with 11 9% of doses. Oxygen desaturation 
occurred with 9 8% of doses 

Other reactions during the dosing pro- 
cedure occurred with fewer than 1% of doses 
and included endotracheal tube reflux, pallor, 
vasoconstriction hypotension, endotracheal 
tube blockage, hypertension, hypocarbia, 
hypercarbia, and apnea No deaths occurred 
during the dosing procedure, and all reac- 
tions resolved with symptomatic treatment 

The occurrence of concurrent illnesses 
common in premature infants was evaluated 
in the controlled trials. The rates in all con- 
trolled studies are in Table 3 

All Controlled Studies 


Pulmonary air leaks 

When all controlled studies were pooled, 
there was no difference in intracranial hemor- 
rhage However, in one of the single-dose res- 
cue studies and one of the multiple-dose 
prevention studies, the rate of intracranial 
hemorrhage was significantly higher in 
SURVANTA patients than control patients 
(63.3% v 30.8%. P= 0.001; and 48.8% v 
34.2%, P = 047, respectively). The rate in 
a Treatment IND involving approximately 4400 
infants was lower than in the controlled trials 

In the controlled clinical trials, there was 
no effect of SURVANTA on results of common 
laboratory tests: white blood cell count 
and serum sodium, potassium, bilirubin, 

More than 3700 pretreatment and post- 
treatment serum samples were tested by 
Western Blot immunoassay for antibodies to 
surfactant-associated proteins SP-B and 
SP-C No IgG or IgM antibodies were 

Several other complications are known to 
occur in premature infants The following 
conditions were reported in the controlled 
clinical studies The rates of the complica- 
tions were not different in treated and control 
infants, and none of the complications were 
attributed to SURVANTA. 
Respiratory: lung consolidation, blood from 
the endotracheal tube, deterioration after 
weaning, respiratory decompensation, sub- 
glottic stenosis, paralyzed diaphragm, respi- 
ratory failure. 

Cardiovascular: hypotension, hypertension, 
tachycardia, ventricular tachycardia, aortic 
thrombosis, cardiac failure, cardio- 
respiratory arrest, increased apical pulse, 
persistent fetal circulation, air embolism, total 
anomalous pulmonary venous return. 
Gastrointestinal: abdominal distention, hem- 
orrhage, intestinal perforations, volvulus, 
bowel infarct, feeding intolerance, hepatic 
failure, stress ulcer 
Renal: renal failure, hematuria. 
Hematologic: coagulopathy, thrombo- 
cytopenia, disseminated intravascular 

Central Nervous System: seizures 
Endocrine/Metabolic: adrenal hemorrhage, 
inappropriate ADH secretion, hyper- 

Musculoskeletal: inguinal hernia 
Systemic: fever, deterioration. 
Follow-Up Evaluations 
To date, no long-term complications or 
sequelae ot SURVANTA therapy have been 

Single Dose Studies 

Six-month adjusted-age follow-up evaluations 
ot 232 infants (115 treated) demonstrated no 
clinically important differences between 
treatment groups in pulmonary and neu- 
rologic sequelae, incidence or severity of reti- 
nopathy of prematurity, rehospitalizations, 
growth, or allergic manifestations. 
Multiple-Dose Studies 
Six-month adjusted age follow-up evaluations 
have not been completed Preliminarily, in 
605 (333 treated) ot 916 surviving infants, 
there are trends for decreased cerebral palsy 
and need for supplemental oxygen in 
SURVANTA infants Wheezing at the time ot 
examination tended to be more frequent 
among SURVANTA infants, although there 
was no difference in bronchodilator therapy. 

Twelve-month follow-up data from the mul- 
tiple-dose studies have been completed in 
328 (171 treated) ot 909 surviving infants To 
date no significant differences between treat- 
ments have been found, although there is a 
trend toward less wheezing in SURVANTA 
infants in contrast to the six month results 


Overdosage with SURVANTA has not been 
reported. Based on animal data, overdosage 
might result in acute airway obstruction. 
Treatment should be symptomatic and 

Rales and moist breath sounds can tran- 
siently occur after SURVANTA is given, and 
do not indicate overdosage Endotracheal 
suctioning or other remedial action is not 
required unless clear-cut signs of airway 
obstruction are present. 


SURVANTA (beractant) Intratracheal Suspen- 
sion is supplied in single-use glass vials 
containing 8 mL of SURVANTA (NDC 
0074-1040-08). Each milliliter contains 25 mg 
of phospholipids (200 mg phospholipids/ 
8 mL) suspended in 9% sodium chloride 
solution. The color Is off-white to light brown. 
Store unopened vials at refrigeration tem- 
perature (2-8°C). Protect from light Store 
vials in carton until ready for use. Vials are for 
single use only Upon opening, discard 
unused drug 

June, 1991 


"What are R.T.s Saying 

about Capnography in 

Critical Care?" 

" If we limit the monitor to an alveolar 
gas monitor, we are truly losing out 
on some of the benefits.** 

'* ...physicians are now aware of the 
availability of the monitor and have 
come to expect it...'* 

...The latest scoop on capnography is now 
available on videocassette. And it's yours 
for the askingL^ Qn ^ informational 

videotape, sponsored by 
Ohmeda, you'll heai 
RTs from around 
the country sharing 
L their viewpoints and 
experience with 
capnography in 
critical care. 

For your complimentary copy, 
call Ohmeda Customer Service 
1-800-345-2700 or 

Circle 144 on reader service card 

For your convenience, and direct access, the advertisers in 
this issue and their phone numbers are listed below. Please 
use this directory for requesting written material or for any 
question you may have. 


AARC Information 214-243-2272 

AVL Scientific Corp 

Chad Therapeutics 

CNS Inc 




DHD Medical 


Drager Critical Care Systems 


Invacare Corp 








Ross Laboratories 

Sherwood Medical 

Siemens Medical Systems 

Transtracheal Systems 



a P-value comparing groups in controlled studies 



study of 160 subjects would be need- 
ed to exclude a 10% change in symp- 
tom score. In these patients NCPAP 
did not improve sleeping respiration, 
exercise tolerance, daytime symp- 
toms, or left ventricular function. As 
two of the subjects studied de- 
teriorated on active therapy, NCPAP 
may be detrimental in some cases. 

Survival of Patients with Chronic 
Obstructive Pulmonary Disease 
Receiving Long-Term Domiciliary 
Oxygen Therapy — K Strom. Am 
Rev RespirDis 1993:147:585. 

Previous trials, in which 76 to 100% 
of the patients were men, have 
shown a varied survival time in 
chronic obstructive pulmonary dis- 
ease treated with long-term dom- 
iciliary oxygen therapy. We have an- 
alyzed predictors of survival in 
chronic obstructive pulmonary dis- 
ease, including sex-related differ- 
ences in survival in 403 patients (201 
men) registered in a national register 
when starting long-term oxygen ther- 
apy between January 1, 1987 and 
June 30, 1989. This register covers 
the whole of Sweden with a popula- 
tion of 8.4 million. Some 90%' of all 
the patients receiving long-term oxy- 
gen therapy are included. Lung func- 
tion and performance status pre- 
dicted survival during oxygen ther- 
apy in men, whereas a poor per- 
formance status and the presence of 
orally administered steroid medica- 
tion predicted poor survival in wom- 
en. Oral steroid medication use was 
correlated with an increased mortal- 
ity rate in women (relative risk of 
death, 2.13; 95% confidence interval, 
1.38 to 3.29; p < 0.001) and showed 
no tendency to improve survival in 
men. In patients not receiving oral 
steroids, women had a lower mortal- 
ity rate than did men (relative risk of 
death, 0.58; 95% confidence interval, 
0.39 to 0.97; p < 0.05). Our data in- 
dicate that in patients not receiving 
oral steroid medication, women have 

a better survival than do men. An in- 
creased mortality was found in wom- 
en receiving oral steroid medication 
that might be caused by an increased 
susceptibility to the side effects of 
oral steroids. 

Airway Obstruction and Ventila- 
tion-Perfusion Relationships in 
Acute Severe Asthma — A Ferrer, J 
Roca, PD Wagner, FA Lopez, R 
Rodriguez-Roisin. Am Rev Respir 
Dis 1993:147:579. 

We have investigated the time course 
of ventilation-perfusion (Va/Q) mis- 
match and airflow obstruction in 18 
patients with acute severe asthma 
with the objective to identify poten- 
tial differences according to the clin- 
ical severity of the attacks. Nine pa- 
tients were hospitalized and nine 
were discharged (emergency room 
stay < 24 h) according to the clinical 
criteria of the attending physicians. 
Spirometry and Va/Q inequality (mul- 
tiple inert gas technique) were meas- 
ured within the first 6 h of treatment 
in the emergency room, hospitalized 
patients (in relation to those dis- 
charged) showed lower airflow rates 
(FEV,, 31 ±3 versus 46 ± 6% pre- 
dicted SEM) and greater Va/Q mis- 
match (as assessed by the dispersion 
of blood flow distributions (logSD 
Q) (1.28 ± 0.11 versus 0.92 ±0.11; 
normal values < 0.6). Even though 
the rate of improvement of airflow 
was similar in both groups (without 
returning to normal limits). Va/Q re- 
lationships improved at different 
rates in each group and reached nor- 
mal values at the end of the study. 
Although in hospitalized patients the 
recovery of Va/Q abnormalities was 
delayed in relation to airflow rates, 
the time course in discharged pa- 
tients was the same. Previous studies 
have shown a dissociation between 
spirometry and Va/Q inequality, sug- 
gesting that whereas airflow rates pre- 
dominantly reflect bronchoconstric- 
tion of large airways, Va/Q mismatch 

is more related to obstructive chang- 
es in peripheral airways. Our results 
support this hypothesis and suggest 
that the more severe the asthma at- 
tacks, the more severe the obstruc- 
tive changes involving peripheral air- 
ways for a given degree of wide- 
spread airway narrowing. 

The Effect of Frequency and Mean 
Airway Pressure on Volume De- 
livery during High-Frequency Os- 
cillation — V Chan, A Greenough, 
AD Milner. Pediatr Pulmonol 1993; 

The performance of a commercially 
available oscillator (SensorMedics 
3100) at different frequencies was as- 
sessed. A frequency response curve 
of a pneumotachograph system was 
constructed and this was used to 
measure the volume delivered by the 
oscillator to a lung model. The vol- 
ume delivered by a constant dia- 
phragm displacement was dem- 
onstrated to be inversely proportional 
to the frequency, but unaffected by 
increasing mean airway pressure 
from 15 to 25 cm rLO. The volume 
delivered during high frequency os- 
cillation (HFO) was then assessed in 
8 infants, median gestational age 29 
weeks. The infants were studied at 
two frequencies, 10 and 15 Hz. both 
of which were used at two levels of 
mean airway pressure (MAP): 2 and 
5 cm H2O above the MAP level pre- 
viously used during conventional 
ventilation. The delivered volume was 
not significantly different at the two 
MAP levels, but was significantly 
greater at 10 than 15 Hz at both 
MAP levels (p < 0.03); at MAP +2 
cm H:0 above baseline the reduction 
in delivered volume was from a me- 
dian of 1.54 mL/kg (range. 0.88- 
3.12) at 10 Hz to 1.18 mL/kg (range. 
0.65-4.5) at 15 Hz. These results sug- 
gest that higher frequencies would re- 
quire an increase in the oscillator dis- 
placement if effective gas exchange 
is to be maintained. 



Do livers Oxygen 


Every Breath! 

Pulsair's oxygen conservation device, the POMS 20/50, 

automatically delivers the prescribed amount of 

oxygen during exercise and rest without 

adjustment! Approved for sleep. 

Configurations available for any oxygen source 

(20 psi and 50 psi) including hospital outlets. 

Provides up to 300% more use time. 

Call, write or fax for more information. 

US 1-800-327-031 3 
Fax 407466-4366 
Int'l 407465-6688 

The Pulsair Building 

P.O. Box 2609 

Ft. Pierce, FL 34954 "World Leadership in Liquid 

Oxygen & Conservation Technology" 

Circle 102 on reader service card 

The 1992 bound volume of Respiratory Care is available The 12 issues of Volume 37 
are bound in a blue buckram cover and can be imprinted with your name or the name of 
your institution at no extra charge The cost per bound volume is S40 00 for AARC mem- 
bers (with valid membership number) and $80 00 tor nonmembers Shipping is included 
for customers within the US and Canada 

For a limited time, you may order the 1987, 1988. and 1989 bound volumes al the dis- 
count price of $30 00 tor AARC members and $70 00 for nonmembers The 1990 and 
1991 Bound Volumes are now available for $35 00 for AARC members and $75 00 for 
nonmembers Orders must be prepaid, or include an institutional purchase order, or 
charge to a credit card 

□ Please send bound volume tor 1992 at $40/$80 

□ Please send bound volume for 1991 at $35/375 

□ Please send bound volume tor 1990 at $35/$75 

□ Please send bound volume tor 1989 at $30/$75 
S Please send bound volume for 1988 at S30/S70 
3 Please send bound volume for 1987 at S30/S70 

"1 Check enclosed 

3 Purchase order enclosed. 

3 Visa 

□ MasterCard 


Exp date 


AARC Member # 





State Zip 

Name to be imprinted 

Daedalus Ente 

prises Inc 

P0 Box 29B86 

■ Dallas TX 75229 • FAX (214) 484-6010 


Rapid Improvement of Static Com- 
pliance after Surfactant Treatment 
in Preterm Infants with Respira- 
tory Distress Syndrome — E Ba- 

raldi. A Pettenazzo. M Filippone. GP 
Magagnin. OS Saia, F Zacchello. Pe- 
diatr Pulmonol 1993:15:157. 

Respiratory mechanics were meas- 
ured in 20 preterm infants before and 
in the 24-hr period after treatment 
with surfactant. All infants were en- 
rolled in the rescue clinical trial with 
Curosurf® carried out in the Neo- 
natal Intensive Care Unit. They re- 
ceived a dose of 200 mg/kg lipid sur- 
factant intratracheally after birth. 
Static compliance of the respiratory 
system (C rs ) was measured by the 
single breath occlusion technique 
during both spontaneous and me- 
chanical ventilation. Resistance of 
the respiratory system (R rs ) and ex- 
piratory time constant (T rs ) were also 
measured. As early as 3 hr after sur- 
factant administration, a significant 
improvement of 45% in C rs meas- 
ured during mechanical ventilation 
(C n V) was noted (0.40 ± 0.14 vs 
0.58 ±0.17 mL/cm fTO/kg, p < 
0.001). together with a significant 
improvement of the arterial/alveolar 
O: tension ratio [Ps/aoJ (0.12 ± 0.03 
vs 0.30 ± 0.16. p < 0.01). The im- 
provement of C rs V and Pa/AO : was 
confirmed 24 hr later (0.55 ± 0.15 
mL/cm H : 07kg and 0.33 ±0.18. re- 
spectively). A significant correlation 
was found between C rs and Pa/AO : (•' 
= 0.56. p < 0.001). Time constant 
values were significantly higher after 
surfactant treatment (0.15 ± 0.07 vs 
0.09 ± 0.03 sec: p < 0.01). R„ re- 
mained unchanged. These data in- 
dicate that Curosurf® given intra- 
tracheally after birth determines a 
rapid improvement of respiratory 
mechanics as soon ?.s 3 hr after dos- 
ing, together with the improvement 
of oxygenation. From the findings 




obtained with the present study, we 
show evidence that respiratory sys- 
tem mechanics may be a useful phys- 
iological measure to guide ventila- 
tory strategy following surfactant 

Episodes of Spontaneous Desatur- 
ations in Infants with Chronic 
Lung Disease at Two Different 
Levels of Oxygenation — C McE- 
voy, M Durand, V Hewlett. Pediatr 
Pulmonol 1993;15:140. 

The optimal range of pulse oximeter 
oxygen saturation (S a O:) for infants 
with chronic lung disease (CLD) has 
not been well established. We quan- 
tified episodes of spontaneous de- 
saturation, at two different ranges of 
S a c>2- For 1 hr each, we alternatively 
administered inspired O: concentra- 
tions (Fio;) necessary to maintain an 
S a0 ; of 94-96% or 87-91% to 21 pa- 
tients (mean birth weight, 865 g; ges- 
tational age, 27.3 weeks; postnatal 
age 40.6 days) with CLD (defined by 
Fio: > 0.21 at 28 days and radio- 
graphic evidence). S a o 2 was mon- 
itored with the Nellcor N-200 ox- 
imeter and analyzed by a computer 
program (SatMaster). The percentage 
of time the infants desaturated to lev- 
els of SaO: < 85 and < 80% revealed 
significantly fewer spontaneous epi- 
sodes during the hour of higher base- 
line SaO; (p < 0.0002). Comparison 
of episodes of spontaneous desatura- 
tion to SaO: < 80 and < 85%, lasting 
0-15, 16-30, 31-45 sec also showed 
significant differences between the 
two levels of S a O;- We conclude that 
when infants with CLD are main- 
tained at a higher S a o ; , they probably 
experience fewer episodes of spon- 
taneous desaturations, because of 
less alveolar hypoxia. We believe 
that attempts at weaning the Fio; 
should be tempered with the need of 
maintaining an adequate S a o:- There- 
fore, prolonged monitoring of oxy- 

genation in infants with CLD at dif- 
ferent levels of SaO: could be helpful 
during the weaning process. 

Sensitization and Exposure to In- 
door Allergens as Risk Factors for 
Asthma among Patients Presenting 
to Hospital — LE Gelber, LH Seltzer. 
JK Bouzoukis. SM Pollart, MD 
Chapman, TAE Platts-Mills. Am 
RevRespirDis 1993;147:573. 

To investigate the role of indoor al- 
lergens in adult patients with acute 
asthma, we conducted a case-con- 
trolled study on patients presenting 
to an emergency room. One hundred 
and fourteen patients and 1 14 control 
subjects were enrolled over a 1-yr 
period in Wilmington, Delaware. 
Sera were assayed for total IgE, and 
for IgE antibodies to dust mites, cat 
dander, cockroach, grass pollen, and 
ragweed pollen. Dust was obtained 
from 1 86 homes and assayed for dust 
mite, cat, and cockroach allergens. 
IgE antibodies to mite, cat, and cock- 
roach were each significantly as- 
sociated with asthma, and this as- 
sociation was very strong among 
participants without medical insur- 
ance and among African Americans. 
Among 99 uninsured participants, 
sensitization to one of the indoor al- 
lergens (> 200 RAST units) was 
present in 28 of 57 asthmatics and in 
one of 42 control subjects (odds ra- 
tio, 39; confidence interval, 9.4 to 
166). For cat and cockroach, the 
combination of sensitization and 
presence of allergen in the house was 
significantly associated with asthma. 
Furthermore, there was a strong in- 
verse relationship between IgE anti- 
bodies to cat and to cockroach, and 
the risk of this sensitization was in 
large part restricted to homes or ar- 
eas with high levels of allergen. Thir- 
ty-eight percent of the asthmatics, 
but only 8% of the control subjects, 
were allergic to one of the three in- 
door allergens, and had high levels 

of the relevant allergen in their hous- 
es (odds ratio. 7.4; confidence inter- 
val, 3.3 to 16.5). The results suggest 
that the risk for asthma related to 
sensitization to indoor allergens ap- 
plies to a large proportion of adults 
with acute asthma and that this risk 
is prominent among the socioeco- 
nomic groups that have suffered the 
largest increase in both morbidity 
and mortality from asthma. 

Collateral Ventilation and Gas Ex- 
change during Airway Occlusion 
in the Normal Human Lung — NW 

Morrell, CM Roberts. T Biggs. WA 
Seed. Am Rev Respir Dis 1993: 


The effectiveness of collateral ven- 
tilation in maintaining alveolar gas 
tensions in obstructed lung segments 
was investigated using fiberoptic 
bronchoscopy to place an occluding 
catheter-tip balloon in selected lobar 
and segmental bronchi in supine nor- 
mal human subjects. Gas tensions 
from beyond the occlusion were 
measured with a respiratory mass 
spectrometer. Collateral ventilation 
is known to be minimal between 
lobes; therefore, values measured in 
obstructed lobes provide a control. 
No significant difference was found 
between the partial pressures of oxy- 
gen or carbon dioxide measured in 
obstructed lobes and in obstructed 
segments. In both cases respiratory 
gas tensions approached reported 
values for mixed venous levels. The 
time taken to attain a steady state of 
gas composition in the obstructed 
lung was rapid (approximately 50 s), 
and it was no different for lobes and 
segments. In addition, collateral ven- 
tilation was assessed by measuring 
the amount of helium reaching oc- 
cluded lobes and segments when 
subjects breathed a mixture of 21% 
oxygen and 79% helium. The rate of 
rise in helium concentration was less 
than 1%/min in both lobes and seg- 




merits, a figure that may be ex- 
plained by delivery of helium in re- 
circulated blood rather than by col- 
lateral ventilation. We conclude that 
intersegmental collateral ventilation 
has a negligible role in the main- 
tenance of alveolar gas tensions in 
supine normal humans during tidal 

Inhaled Nitric Oxide Selectively 
Reverses Human Hypoxic Pul- 
monary Vasoconstriction without 
Causing Systemic Vasodilation — 

CG Frostell. H Blomqvist. G He- 
denstierna, J Lundberg, WM Zapol. 
Anesthesiology 1993;78:427. 

BACKGROUND: Nitric oxide (NO), 
an endothelium-derived relaxing fac- 
tor, acts as a local vasodilator. The 
authors examined the effects of NO 
on pulmonary and systemic circula- 
tion in human volunteers. METH- 
ODS: Nine healthy adults were stud- 
ied awake while breathing ( 1 ) air, (2) 
12% O: in N : , (3) followed by the 
same mixture of O: and N : con- 
taining 40 ppm of NO. Pulmonary 
artery and radial artery pressures 
were monitored. RESULTS: The 
PuO; decreased from 106 ± 4 (mean ± 
SEM) while breathing air (21% 2 ) 
to 47 ± 2 torr after 6 min of breath- 
ing 12% 2 . Concomitantly, the pul- 
monary artery mean pressure (PAP) 
increased from 14.7 ± 0.8 torr to 
19.8 ± 0.9 torr, and the cardiac out- 
put (CO.) increased from 6.1 ± 0.4 
to 7.7 ± 0.6 L/min. After adding 40 
ppm NO to the inspired gas while 
maintaining the F102 at 0.12, the PAP 
decreased (p < 0.01, by analysis of 
variance) to the level when breathing 
air while the P a o : and P a co: were un- 
changed. The dilation (or recruit- 
ment) of pulmonary vessels produced 
by inhaling NO during hypoxia was 
not accompanied by any alteration in 
the systemic vascular resistance or 
mean arterial pressure (MAP). The 
authors also examined the effects of 
inhaling NO while breathins air. 

Breathing 40 ppm NO in 21% : for 
6 min produced no significant chang- 
es of PAP, CO.. Pao 2 , MAP, or cen- 
tral venous pressure. Plasma endo- 
thelinlike immunoreactivity concen- 
trations did not change either during 
hypoxia or hypoxia with NO inhala- 
tion. CONCLUSIONS: Inhalation of 
40 ppm NO selectively induced pul- 
monary vasodilation and reversed 
hypoxic pulmonary vasoconstriction 
in healthy humans without causing 
systemic vasodilation. 

Randomized Evaluation of Pulse 
Oximetry in 20,802 Patients: I. De- 
sign, Demography, Pulse Oximetry 
Failure Rate, and Overall Com- 
plication Rate — JT Moller, T Peder- 
sen, LS Rasmussen, PF Jensen, BD 
Pedersen, O Ravlo, et al. Anes- 
thesiology 1993;78:436. 

BACKGROUND: Although pulse 
oximetry is currently in widespread 
use, there are few data documenting 
improvement in patient outcome as a 
result of the use of oximetry. The au- 
thors describe the study design, pa- 
tient demographic findings, data val- 
idation, pulse oximetry failure rate, 
and overall postoperative complica- 
tion rates in the first large pros- 
pective randomized multicenter clin- 
ical trial on perioperative pulse 
oximetry monitoring. METHODS: In 
five Danish hospitals, by random as- 
signment, monitoring did or did not 
include pulse oximetry for patients 
18 yr of age and older, whether 
scheduled for elective or emergency 
operations, or for regional or general 
anesthesia, except during cardiac and 
neurosurgical procedures. Operation- 
al definitions were established for 
perioperative events and postopera- 
tive complications. The data were 
collected preoperatively, during an- 
esthesia, in the postanesthesia care 
unit, and until the day of discharge 
from the hospital or the seventh post- 
operative day. RESULTS: Of 20,802 
patients, 10,312 were assigned to the 

oximetry group and 10,490 to the 
control group. In general, the dem- 
ographic data, patient factors, and 
anesthetic agents used were distrib- 
uted evenly. A slight intergroup dif- 
ference was found in the distribution 
of age, duration of surgery, some 
types of surgery, and some types of 
anesthesia. The total failure rate of 
the oximetry was 2.5%, but it in- 
creased to 7.2% in patients with 
American Society of Anesthesiolo- 
gists physical status 4 (p < 0.00001). 
In 14.9% of the patients, one or more 
events occurred in the operating 
room and 13.5% in the post- 
anesthesia care unit. The overall 
postoperative complication rate was 
9.7%. The total rates of cardio- 
vascular and respiratory complica- 
tions were 2.78% and 3.50%, re- 
spectively. Within the first seven 
postoperative days, 0.47% of the pa- 
tients died. Anesthesia was not 
thought to have been solely re- 
sponsible for any death, but in 7 pa- 
tients (1 per 3,365), it was a possible 
contributory factor. CONCLUSIONS: 
Despite the finding of a few sig- 
nificant intergroup differences, the 
randomization was well balanced 
with a high validity of data. The 
overall postoperative complication 
rate was similar to that in other re- 
cent morbidity and mortality studies. 

Deposition in Asthmatics of Par- 
ticles Inhaled in Air or in Helium- 
Oxygen — M Anderson, M Svarten- 
gren, G Bylin, K Philipson, P Cam- 
ner. Am Rev Respir Dis 1993:147: 

Ten subjects with asthma inhaled 3.6 
//m particles labeled with '"In in air 
and in a helium-oxygen mixture (He- 
: ) at 0.5 and at 1.2 L/s. Lung re- 
tention was measured after zero and 
after 24 h, and the percentage 24-h 
retention (Ret; 4 ) was taken to repre- 
sent the fraction deposited in the al- 
veolar part of the lung. For both in- 
halation rates, Ret:4 was significantly 





. , 


We not only meet all NAEP standards - 
weVe set a few of our own. 

Our patented flow-sampling technology, 
for example, that reduces wear and delivers 
superior accuracy and reproducibility. 1,2 

Our easy-to-use, easy-to-read, easy-to-clean 
(but hard-to-wear-out) design, for another. 

Our meticulous quality-control testing of 
every unit. 

And our unmatched program of professional 
and patient support-including the industry's 
only comprehensive peak flow monitoring 

All of which helps explain why physicians 
have made us the standard - America's #1 peak 
flow meter in the hospital, office, and home. 

Let us show you how we can meet your 
tough standards. Call us toll-free at 

Peak Flow 


60 to 880 L/min 30 to 390 L/min 

Setting the standard for peak flow monitoring. 

HealthScan Products Inc., 908 Pompton Avenue, Cedar Grove, NJ 07009-1292 

References: 1. Data on tile, HealthScan Products Inc. 2. Shapiro S, Hendler J, Ogirala R, et al.: An evaluation of the accuracy of Assess and MiniWright peak flowmeters. Chest 99:358-362, 1991. 
ASSESS Standard and Low Range meters meet National Asthma Education Program Technical Standards for Peak Flow Meters, January, 1991. AA71 0002-0 5/92 

Circle 136 on reader service card 

Make work of breathing light as a feather! 

Drager is the first to offer you Minimum Pressure Ventilation, a feather light synchronizing 
ventilation philosophy that lets you deliver volume constant pressure control. 

Up until now, you've been left literally in the dark when it comes to knowing how your patients 
are responding to respiratory therapy. Evita's integrated screen lets you view analogue 
ventilation waveforms, plain text, status information and alarm messages. 

And what's more, Drager has streamlined operations, making it easier for you to fine tune 
ventilation modes, including Airway Pressure Release Ventilation (APRV), a unique ventila- 
tion mode to treat ARDS patients. All of this adds up to greater patient safety, maximum 
patient comfort and ultimately, your peace of mind. Your know how and Drager technology -- a 
powerful partnership for saving lives For the difference your patients can feel, 

Drager : Technology for Life Circle 101 on reader service card 

4101 Pleasant Valley Drive Suite 1 00 Chantilly.VA 22021 Tel (703)817-0100 Fax:(703)817-0101 



higher when particles were inhaled 
with He-O: than with air. The in- 
crease in Ret;a seemed to be larger in 
subjects with asthma than in healthy 
persons earlier studied. Ret:a was 
correlated with changes in both large 
and small airways, especially when 
the particles were inhaled with He- 
O;. Our data suggest that inhalation 
of drugs in He-0 : might be of ther- 
apeutic value when treating patients 
with severely obstructed airways. 

The Effect of Varying Levels of 
Outdoor-Air Supply on the Symp- 
toms of Sick Building Syndrome — 

R Menzies, R Tamblyn, J-P Farant. J 
Hanley, F Nunes, R Tamblyn. N 
Engl J Med 1993:328:821. 

BACKGROUND: The sick building 
syndrome is the term given to a con- 
stellation of symptoms reported by 
workers in modern office buildings, 
hypothesized to occur when the sup- 
ply of outdoor air is reduced, be- 
cause of the accumulation of con- 
taminants arising from within the 
building. We undertook this study to 
determine the effect of changing the 
supply of outdoor air in four office 
buildings on the symptoms reported 
by workers and their perception of 
the indoor environment. METHODS: 
Within each of three consecutive 2- 
week blocks, the ventilation systems 
in each building were manipulated, 
in random order, to deliver to the in- 
door environment an intended 20 or 
50 ft- (0.57 or 1.4 m') of outdoor air 
per minute per person for 1 week at a 
time. Each week, the participants, 
unaware of the experimental inter- 
vention, reported symptoms and the 
indoor environment was thoroughly 
evaluated. RESULTS: Of 1.838 el- 
igible workers in the four buildings, 
1,546 (84%) participated in the 
study. The supply of outdoor air av- 
eraged 7% and 32% in the ventila- 
tion systems and 30 and 64 ft' (0.85 

and 1.8 m 3 ) per minute per person in 
the work sites at the lower and high- 
er ventilation levels, respectively. 
These changes in the supply of out- 
door air were not associated with 
changes in the participants* ratings 
of the office environment or in symp- 
tom frequency (crude odds ratio, 1.0; 
95% confidence interval, 0.9 to 1.1). 
After work-site measures of ventila- 
tion, temperature, humidity, and air 
velocity were included in the re- 
gression analysis, the adjusted odds 
ratio was also 1.0 (95% confidence 
interval, 0.8 to 1.2). CONCLU- 
SIONS: Increases in the supply of 
outdoor air did not appear to affect 
workers' perceptions of their office 
environment or their reporting of 
symptoms considered typical of the 
sick building syndrome. 

Randomized Evaluation of Pulse 
Oximetry in 20,802 Patients: II. 
Perioperative Events and Post- 
operative Complications — JT Mol- 
ler, NW Johannessen, K Espersen, O 
Ravlo, BD Pedersen, PF Jensen, et 
al. Anesthesiology 1993;78:445. 

BACKGROUND: The authors de- 
scribe the effect of pulse oximetry 
monitoring on the frequency of un- 
anticipated perioperative events, 
changes in patient care, and the rate 
of postoperative complications in a 
prospective randomized study. 
METHODS: The study included 
20,802 surgical patients in Denmark 
randomly assigned to be monitored 
or not with pulse oximetry in the op- 
erating room (OR) and post- 
anesthesia care unit (PACU). RE- 
SULTS: During anesthesia and in the 
PACU. significantly more patients in 
the oximetry group had at least one 
respiratory event than did the control 
patients. This was the result of a 19- 
fold increase in the incidence of di- 
agnosed hypoxemia in the oximetry 
group than in the control group in 

both the OR and PACU (p < 
0.00001). In the OR, cardiovascular 
events were observed in a similar 
number of patients in both groups, 
except myocardial ischemia (as de- 
fined by angina or ST-segment de- 
pression), which was detected in 12 
patients in the oximetry group and in 
26 patients in the control group (p < 
0.03). Several changes in PACU care 
were observed in association with 
the use of pulse oximetry. These in- 
cluded higher flowrate of supple- 
mental oxygen (p < 0.00001). in- 
creased use of supplemental oxygen 
at discharge (p < 0.00001), and in- 
creased use of naloxone (p < 0.02). 
The rate of changes in patient care as 
a consequence of the oximetry mon- 
itoring increased as the American 
Society of Anesthesiologists physical 
status worsened (p < 0.00001). One 
or more postoperative complications 
occurred in 10% of the patients in 
the oximetry group and in 9.4% in 
the control group (difference not sig- 
nificant). The two groups did not dif- 
fer significantly in cardiovascular, 
respiratory, neurologic, or infectious 
complications. The duration of hos- 
pital stay was a median of 5 days in 
both groups (difference not signif- 
icant). An equal number of inhospital 
deaths were registered in the two 
groups. Questionnaires, completed 
by the anesthesiologists at the five 
participating departments, revealed 
that 1 8% of the anesthesiologists had 
experienced a situation in which a 
pulse oximeter helped to avoid a se- 
rious event or complication and that 
80% of the anesthesiologists felt 
more secure when they used a pulse 
oximeter. CONCLUSIONS: This 
study demonstrated that pulse oxim- 
etry can improve the anesthesiol- 
ogist's ability to detect hypoxemia 
and related events in the OR and 
PACU and that the use of the ox- 
imeter was associated with a sig- 
nificant decrease in the rate of myo- 
cardial ischemia. Although moni- 
toring with pulse oximetry prompted 




a number of changes in patient care, 
a reduction in the overall rate of 
postoperative complications was not 

The Occurrence of Sleep-Disord- 
ered Breathing among Middle- 
Aged Adults— T Young, M Palta, J 
Dempsey, J Skatrud, S Weber, S 
Badr. N Engl J Med 1993;328: 1230- 

BACKGROUND: Limited data have 
suggested that sleep-disordered 
breathing, a condition of repeated ep- 
isodes of apnea and hypopnea during 
sleep, is prevalent among adults. 
Data from the Wisconsin Sleep Co- 
hort Study, a longitudinal study of 
the natural history of cardiopulmo- 
nary disorders of sleep, were used to 
estimate the prevalence of undiag- 
nosed sleep-disordered breathing 
among adults and address its im- 
portance to the public health. METH- 
ODS: A random sample of 602 em- 
ployed men and women 30 to 60 
years old were studied by overnight 
polysomnography to determine the 
frequency of episodes of apnea and 
hypopnea per hour of sleep (the ap- 
nea-hypopnea score). We measured 
the age- and sex-specific prevalence 
of sleep-disordered breathing in this 
group using three cutoff points for 
the apnea-hypopnea score (5. 10. 
and 15); we used logistic re- 
gression to investigate risk factors. 
RESULTS: The estimated preva- 
lence of sleep-disordered breathing, 
defined as an apnea-hypopnea score 
of 5 or higher, was 9% for women 
and 24% for men. We estimated that 
2% of women and 4% of men in the 
middle-aged work force meet the 
minimal diagnostic criteria for the 
sleep apnea syndrome (an apnea- 
hypopnea score of 5 or higher and 
daytime hypersomnolence). Male sex 
and obesity were strongly associated 
with the presence of sleep-disordered 
breathing. Habitual snorers, both men 

and women, tended to have a higher 
prevalence of apnea-hypopnea scores 
of 15 or higher. CONCLUSIONS: 
The prevalence of undiagnosed sleep- 
disordered breathing is high among 
men and is much higher than pre- 
viously suspected among women. 
Undiagnosed sleep-disordered breath- 
ing is associated with daytime hyper- 

Nonbronchoscopic Bronchoalveo- 
lar Lavage in Mechanically Venti- 
lated Infants: Technique, Efficacy, 
and Applications — AC Koumbour- 
lis, G Kurland. Pediatr Pulmonol 

Bronchoalveolar lavage with the fib- 
eroptic bronchoscope is commonly 
used for the diagnosis of pulmonary 
infections in mechanically ventilated 
adults and children. However, its use 
for intubated infants is precluded be- 
cause the small artificial airway does 
not permit the passage of the bron- 
choscope. We have developed a tech- 
nique for nonbronchoscopic bron- 
choalveolar lavage, performed via a 
sterile, disposable feeding tube. We 
have used this technique in 15 in- 
fants with diffuse interstitial disease 
and/or atelectasis, while they were 
intubated and mechanically ventilat- 
ed. The volume of the lavage efflu- 
ent averaged 70.3% of the volume 
instilled. Specific diagnosis on the 
basis of the cytologic evaluation and/ 
or culture of the lavage fluid was 
possible in 9 (60%) patients. Two pa- 
tients with atelectasis showed radio- 
graphic evidence of improvement fol- 
lowing the procedure. There were no 
complications. We conclude that non- 
bronchoscopic bronchoalveolar lav- 
age is well tolerated, and clinically 
useful in small, mechanically venti- 
lated infants with respiratory failure 
due to diffuse pulmonary disease. 
This technique provides a lower risk 
alternative to more invasive, and 
costly procedures. 

Volume Recruitment Maneuvers 
Are Less Deleterious than Persist- 
ent Low Lung Volumes in the Ate- 
lectasis-Prone Rabbit Lung during 
High-Frequency Oscillation — DM 
Bond. AB Froese. Crit Care Med 

OBJECTIVES: To test whether the 
pulmonary risk of repeated volume 
recruitment is greater or less than the 
risk associated with unreversed ate- 
lectasis during 6 hrs of high-fre- 
quency oscillatory ventilation in the 
atelectasis-prone rabbit lung. DE- 
SIGN: Prospective, controlled, ran- 
domized comparison over 6 hrs of 
ventilator management. SETTING: 
Laboratory. SUBJECTS: Twenty 
eight adult New Zealand white male 
rabbits (weight 2.3 to 2.8 kg). 
BACKGROUND: Controversy ex- 
ists over whether high-frequency os- 
cillatory ventilation should be used 
with volume recruitment maneuvers 
in the atelectasis-prone lung, or be 
used at low mean and peak pressures 
without volume recruitment to avoid 
the risks of even transient pulmonary 
overdistention. Potential risks and 
benefits accompany both alterna- 
tives. INTERVENTIONS: We evalu- 
ated the pulmonary effects of three 
high-frequency oscillatory ventila- 
tion protocols in anesthetized rabbits 
made surfactant deficient by saline 
lavage, using animals ventilated with 
conventional positive-pressure venti- 
lation with positive end-expiratory 
pressure as a reference group; n = 5 
in each group. The three high-fre- 
quency oscillatory ventilation groups 
were ventilated for 6 hrs at 15 Hz 
(900 breaths/min). Fio: = 1 .0. The re- 
peated stretch group received 15-sec 
sustained inflations at 30 cm H : 
mean airway pressure every 20 mins, 
with maintenance mean airway pres- 
sure sufficient to keep P,,o: > 350 ton 
(46.7 kPa). The repeated deflation 
group was maintained at levels that 
produced P a o 2 70 to 120 torr (9.3 to 
16 kPa), with the endotracheal tube 




To place recruitment adver- 
tising, contact Valley Forge 
Press at (800) 220-4979. Ads 
can be faxed to (215) 935- 
8208 or mailed to Respira- 
tory Care, 1288 Valley Forge 
Road, Suite 50, P.O. Box 
1135, Valley Forge, PA 
19482. ■ 




(214) 243-2272 

For a 





L L 

Respiratory Care 

. Supervisor ^ 

Keth Israel is recognized around the world 
as a leading innovator in patient care and 
research and is still setting the pace on many 
medical fronts. We are offering you the 
opportunity to grow, learn and achieve your 
careers' full potential, while encouraging 
others to do the same. 
We are seeking a dedicated and dynamic 
individual who shares our commitment to 
progressive respiratory care of patients in 
both general and critical care areas. You 
will help supervise all aspects of respiratory 
care in our active emergency unitArauma 
center, our four adult ICUs and our general 
nursing unit beds. You will be responsible 
for the direct supervision and guidance of 24 
respiratory therapists, and will work collabo- 
ratively with physicians, nurses and other 
members of the healthcare team. 
As a respiratory care supervisor at Beth 
Israel, you'll have extensive educational 
opportunities including outside conferences, 
and our own ACLS program. 
Ideal candidate must have at least 2 years of 
recent supervisory experience, have NBRC 
registration, and a Mass. license to practice. 
Must have excellent organizational and 
communication skills. 
For the above position, please send resume 
to: Kajal Sen Gupta, Human Resources, 
Beth Israel Hospital, 330 Brookline 
Avenue, Boston, MA 02215. If you are 
interested in other employment opportuni- 
ties, or would like to complete an applica- 
tion, please visit our Employment Office at 
132 Brookline Avenue, 2nd Floor. An 
Equal Opportunity Employer M/F. 

^syjF Beth Israel 
Ijjl Hospital Boston 

^2m A Partnership of Ideas. 




MEDICAL/Respiratory Therapist 




Georgia Baptist Medical Center, a dynamically 
growing Medical center, is seeking a highly 
motivated and skilled practitioner to join the 
Respiratory Care Team as our Clinical 
Respiratory Specialist. This challenging 
position offers an experienced Practitioner the 
ability to serve as a resource for complex patient 
care situations, a liaison between patients, 
nurses and physicians, serve as a consultant to 
any of the health care professionals and 
participate in the professional development of 
the hospital staff. The successful candidate will 
possess a BS degree (BSRT preferred), RRT, 
State licensure, minimum of 5 years, minimum 
of 3 years Neonatal ICU experience and clinical 
instructor experience. Strong communication 
skills are essential. Preferred requirements are 
NRP and PALS certifications, research 
experience and personal computer experience. 

Excellent salary and benefits including medical 
and dental insurance, free life insurance, tuition 
reimbursement, free parking, credit union, 
childcare center, fitness center and much more. 
Applications may be placed in Human resource 
dept., Mon-Wed, 9 - 1 lam and l-3pm, or send 
resume to: 


300 Boulevard NE 

Box 405 
Atlanta, GA 30312 

Equal Opportunity Employer M/F 

1901* 1991 
Gtotnt Btftst 








Enjoy year-round sunshine and gentle breezes ... and 
a healthcare community dedicated to a mission of 
healing ... with St. Anthony's Hospital in St. 
Petersburg, Florida. 

Respiratory TTierapist Supervisor 
1st Shift 

Level 3 Therapists 

3-11 and Pool 

Critical care experience required 

Must be licensed and certified, registry preferred. 
Very competitive salaries and benefits. For more 
information, please contact Carla Gurr, Recruiter, at 
1-800-876-3539 or send resume to: St. Anthony's 
Hospital, Human Resources, 1200 7th Avenue N., St. 
Petersburg, FL 33705. EOE. 


St. Anthony's Hospital 

^1200 7th Avenue N. ■ St. Petersburg, FL 33705 IQlJ 


... — THERAPY — _. 

Immediate supervisory position available at Lake- 
land Regional Medical Center, a 650+ bed acute 
care facility in Central Florida. Responsibilities 
include the daily operation of the department and 
supervising a staff of 55. 

Qualified candidate must be RRT or CRTT with 1 - 
2 years supervisory experience, and a minimum of 
6 years critical care experience. 
Lakeland Regional Medical Center employees 
enjoy up to 23 paid days off annually, competitive 
salary, flexible benefits package including medical 
and dental insurance at pretax rates, and free 
parking. For more information, call (813)687- 
1377 collect or send resume to: 

i Lakeland . 
= == Regional 
^Medical Center 

Employment Services 
P.O. Box 95448 • Lakeland, FL 33804 

equal opportunity employer 

SVJ\^ 6 



CA*- L 







Career Opportunities 

Children's Medical Center of Dallas, a national leader in pediatric treatment and research, is 
dedicated to making life better for children. COME GROW WITH US as we expand to meet 
the challenges and demands for pediatric respiratory care. We are currently seeking 
candidates for the following full-time positions within our Respiratory Care Department: 
Critical Care Therapist Educator 

RRT, current BCLS Instructor certification and /or PALS certification preferred. Two years 
of critical care and l year of general pediatric experience are necessary. Responsibilities 
include orientation/education of departmental staff in critical care areas and other disciplines. 
Respiratory Therapists 

Emergency Center - RRT or registry eligible, pediatric and /or emergency room experience 
a plus. 

General Care-RRT or registry eligible * PICU - RRT with critical care experience. 

All positions require a' current Texas license. For more information and appointment, call 
D.C. Mangum at 214-640-2543 or 1-800-852-0982 or send your resume to: Children's 
Medical Center of Dallas, Attn: Human Resources Dept. 1935 Motor St., Dallas, TX 75235. 




1935 M«or Street - Dallas. Texas 7b?35 • (2w) 920-2000 

The Cure for the Common Career 
Equal Opportunity Employer M/F/D 



Director, Respiratory Therapy 
- Sought by North Mississippi 
Medical Center (NMMC) 
NMMC is a 650 bed, not-for 

Firofit regional referral center, 
ndividual will supervise 40+ 
FTE's and seven supervisors 
with a total budget of $3.4M. 
Preference to those applicants 
familiar with total quality 
management. Requires a 
Registered Respiratory 
Therapist Degree or an 
appropriate Bachelor's Degree 
in Healthcare Management. 
Minimum of three years 
experience within a complex 
hospital system or a large 
stand alone hospital. Contact 
John J. Baumann, Vice 
President, J.J. & H., Ltd., 1785 
The Exchange, Suite 320, 
Atlanta, GA 30339, or call 404- 
952-3877. EOE 



With a 526-bed teaching and referral 
hospital committed to the highest 
level of patient care, Memorial Health 
System is the premier health care 
provider in its region. 


needed to complete our highly 
motivated and skilled staff of 

• Completion of two-year Associate 
Degree in Respiratory Therapy 

• License eligible in Indiana 

For further consideration contact 
Carol Lyle-Ford, Human 
Resources, (219) 284-3274, 
BEND, 615 N. Michigan, South 
Bend, IN 46601. 

Equal Opportunity Employer. 


Health System 







Redefine Home Healthcare in the fresh 
and exciting environment of Alaska. 

Experience a unique spirit of independence 
and interaction with both physicians and 
customers/patients with Abbey Home 
Healthcare, a leader in clinical services. 
We are redefining an industry with our 
focus on cost containment and quality medi- 
cal services for the fastest-growing health- 
care segment today, healthcare at home. 

Enjoying extensive travel opportunities, you 
will help ensure that our patients/customers 
make a successful transition from hospital to 
home care, providing patient and caregiver 
education and instruction in the uses of our 
products and services. You will also conduct 
ongoing follow-up visits as an innovative 
patient care advocate. 

Alaska awaits for those who are people- 
oriented and certified or registered with at 
least 2 years of recent acute care experience 
in home healthcare. A self-motivated, take- 
charge attitude is preferred. 

Surround yourself with a fresh, new envi- 
ronment and build your skill with an entre- 
preneurial focus by working among the best 
in the field. We offer a competitive salary 
and comprehensive benefits. To apply, send 
your resume and salary history to: Kevin 
Dee, Branch Manager, Abbey Home 
Healthcare, 4640 Old Seward Hwy., 
#102, Anchorage, AK 99503. 

Caring fir America's Health at Home 

To place 
an ad 

(215) 935-3301 







S t udy 


NBRC Entry Level Examination 
tutorials are specially written to help 
you pass two' of the most difficult 
sections of the examination Uses 
questions now retired from the actual 
examination question pool (used with 
permission of Applied Measurement 
Professionals, Inc., a wholly owned 
subsidiary of the NBRC). 

Tutorial I: Detecting Equipment 
Malfunctions, Item CD4 — $39.95 
Ea. (AARC member $29.95), add 
$4.50 for shipping and handling 

Tutorial II: Evaluate & Monitor Patient's 
Response to Respiratory Care, 
Asses Patient Response, Item CD5 
— $39.95 Ea, (AARC member 
$29.95), add $4.50 for shipping and 

Order both tutorials and save, Item 
CD6 — $75 Set (member $55), add 
$5.50 for shipping and handling. 

To Order Call 

(214)243-2272 or Fax to 






for the 

Respiratory Care 


Covers four major 
aspects of smoking 
cessation — the impact 
of smoking on illness 
and mortality, 

components, current 
cessation programs, 
and effective 
intervention. A one 
hour videotape lecture 
by Kathleen A. Smalky, 

Item VT35 — $40 Ea 

(AARC Member $35), 
add $4.50 for shipping 
and handling. 

To Order 




Fax to 









"V" ■ ■••■■■ 

■ ■ 



■ , "•* i 

Your company and products gain national attention when featured or mentioned in 
Respiratory Care journal. The journal covers all the facts about the latest technology 
and procedures in respiratory care... and health care buyers know it. 

Your company can benefit from RESPIRATORY Care reprints. 

Here's how — Your products can't get a better recommendation than when they're 
mentioned in Respiratory Care. In our exclusive format, Respiratory Care reprints will prove 
to your company's stockholders, prospective customers, and present customers that you are 
a leader in the industry. And Respiratory Care reprints are great for trade shows. 

Fast, easy, and inexpensive. 

Arrange to have your reprint produced through one convenient department. Reprints are 
printed on gloss stock for immediate impact, and our competitive pricing makes Respiratory 
Care reprints the perfect supplement to your advertising program. Call for a price quote and 
more information on how your company can take advantage of Respiratory Care reprints. 

(214) 243-2272 




For Leaks. 

It Figures. 


t .> i 


Leaks are common with mask ventilation. That's why Respironics' 
BiPAP S/T-D Hospital System is smart enough to calculate when circuit leaks 
occur. And intelligent enough to both tolerate and compensate for them. So, 
in the presence of most leaks, the BiPAP System can deliver pressure at 
prescribed levels. At the same time, it analyzes flow patterns during every m 
breath and makes appropriate adjustments to offer patient sensitive therapy. V-v 

With unsurpassed pressure stability and breath-to-breath sensitivity, the 
BiPAP pressure support ventilator is proving to be effective for appropriate 
candidates, including patients with alveolar hypoventilation, persistent hypoxemia 
and/or hypercapnia, ventilatory muscle dysfunction, post-extubation difficulty and 
upper airway obstruction. 

To find out how the BiPAP S/T-D Hospital System can figure into the needs of 
your facility, just call Respironics at (800) 345-6443. 

The BiPAP SIT-D Hospital System is 
ideal for appropriate candidates in the 
1CU, ER and Recovery Room. 



1001 Murry Ridge Drive, MurrysviUe, Pennsylvania 15668-8550 USA 

Phone: (800) 345-6443 within the USA and Canada 

or (412) 733-0200 Fax:(412)733-0299 

The First Name In Innovative Respiratory Care 

S7 \ l 

The BiPAP System is not intended to provide the total ventilatory requirements of the patient and must not be used as a life support ventilator. 
BiPAP pressure support ventilators are the subject of U.S. Patent #5148802 and other pending U.S. and foreign patents. 

©Respironics Inc. 1993 
Circle 129 on reader service card 

±T^ =f 

It's easy to see how the ACE® 
Aerosol Cloud Enhancer will 
rJd=WaW«i»»lll»»dSlgr.JgMdsl improve your MDI delivery. 
The patented design features a unique shape that matches the aerosol 
plume. This shape allows the ACE to better capture and deliver the 
medication. When the inhaler is actuated, respirable aerosol particles 
are suspended in the ACE for delivery to your patient. Larger, non- 
respirable particles can deposit in the chamber instead of the patient's 
mouth and throat. 

Other features of the ACE include: 

■ Clear Holding Chamber lets you feel confident the inhaler worked 
correctly, and the prescribed dose is available for delivery. 

■ One -Way Valve in the mouthpiece protects the aerosol dose in the 
chamber until inhalation begins. 

■ Coaching Whistle helps to maintain the slow inspiratory flowrate 
recommended for optimum aerosol particle delivery. 

■ Highly Versatile Design allows the ACE to also be used in a vent circuit, 
with an incentive spirometer, or in conjunction with an 
endotracheal airway or resuscitation bag. 

See for yourself how the ACE will help you maximize your MDI deliv- 
ery. For more information on the ACE Aerosol Cloud Enhancer, or the 
entire line of DHD quality respiratory 1 Of\f\ Q/VJ 0000 
care products, call DHD toll-free at J~OUU"Cr±/ "OUUU. 


DIEMOLDING healthcare division 
Canastota, NY 13032 USA Phone (315) 697-2221 Fax (315) 697-8083 

Circle 131 on reader service card 




The Christmas Seal People " 


opened to atmospheric pressure for 
15 sees every 20 mins. Animals in 
the repeated stretch after deflations 
group were managed as in the re- 
peated stretch protocol but each sus- 
tained inflation was preceded by a 
15-sec deflation to functional resid- 
ual capacity. The conventional pos- 
itive-pressure ventilation group was 
ventilated at rates of 30 to 100 
breaths/min. keeping P a o: 70 to 120 
torr (9.3 to 16 kPa). End points in- 
cluded terminal functional residual 
capacity and a compliance index 
computed from respiratory system 
pressure-volume curves. MEASURE- 
6 hrs of ventilation, respiratory sys- 
tem compliance in the repeated 
stretch group had returned to control 
values (1.35 ±0.18 [SD] mL/kg/cm 
FLO). Respiratory system compli- 
ance was significantly less than this 
number in both the repeated de- 
flation (0.89 ± 0.08) and repeated 
stretch after deflations (1.24 ± 0.22) 
groups (p < 0.05). Respiratory sys- 
tem compliance after 3 hrs of con- 
ventional positive-pressure ventila- 
tion decreased to 0.34 ± 0.10 mL/kg/ 
cm FLO. Functional residual capacity 
changes paralleled these changes of 
respiratory system compliance. CON- 
CLUSIONS: These data demonstrate 
that the potential pulmonary risk of 
repeated lung stretch during volume 
recruitment is significantly less than 
the damage arising from the avoid- 
ance of such maneuvers in lungs in 
which alveolar recruitment is pos- 
sible. We conclude that sustained in- 
flations during high-frequency oscil- 
latory ventilation produce the bene- 
fits of increased oygenation for a giv- 
en mean airway pressure plus de- 
creased progression of lung injury. 

Healthcare Reform: The Role of 
Coordinated Critical Care — FB 
Cerra. Crit Care Med 1993:21:457- 




OBJECTIVE: To evaluate and ed- 
itorialize the evolving role of the dis- 
cipline of critical care as a healthcare 
delivery system in the process of 
healthcare reform. DATA SOURC- 
ES: The sources included material 
from the Federal Office of Man- 
agement and Budget. Health Care 
Financing Review, President Bush's 
Office. Association of American 
Medical Colleges, and publications 
of the Society of Critical Care Med- 
were selected that the author felt was 
relevant to the healthcare reform pro- 
cess and its implications for the dis- 
cipline of critical care. DATA EX- 
TRACTION: The data were 
extracted by the author to illustrate 
the forces behind healthcare reform, 
the implications for the practice of 
critical care, and role of critical care 
as a coordinated (managed) care sys- 

tem in the process of healthcare re- 
form. DATA SYNTHESIS: Health- 
care reform has been initiated be- 
cause of a number of considerations 
that arise in evaluating the current 
healthcare delivery system: access, fi- 
nancing, cost, dissatisfactions with 
the mechanisms of delivery, and po- 
litical issues. The reform process will 
occur with or without the in- 
volvement of critical care practi- 
tioners. Reforms may greatly alter the 
delivery of critical care services, ed- 
ucation, training, and research in crit- 
ical care. Critical care has evolved 
into a healthcare delivery system that 
provides services to patients who 
need and request them and provides 
these services in a coordinated (man- 
aged) care model. CONCLUSIONS: 
Critical care practitioners must be- 
come involved in the healthcare re- 
form process, and critical care ser- 

vices that are effective must be pre- 
served, as must the education, train- 
ing, and research programs. Critical 
care as a healthcare delivery system 
utilizing a coordinated (managed) 
care model has the potential to pro- 
vide services to all patients who need 
them and to deliver them in a manner 
that is cost-effective and recognized 
as providing added value. 

Prevention of Hypoxemia during 
Lumbar Puncture in Infancy with 
Preoxygenation — DH Fiser, GA 
Gober, CE Smith, DC Jackson, W 
Walker. Pediatric Emerg Care 1993: 

Hypoxemia has previously been re- 
ported during lumbar puncture (LP) 
in infancy. The purpose of this study 
was to determine whether preoxy- 
genation before the LP would reduce 

The BEST Oxygen Delivery 

• Improved A-a Gradient 

• Reduced Dyspnea 

• Reduced Inspired Minute Ventilation 

• Decreased Work of Breathing 

• Guaranteed 24-hour Compliance 

• Easy to use 

What more do you want? 
If YOU don ? t know about SCOOP® call to 
find out...PERIOD! 

Transtracheal Systems, Inc. 8755 E. Orchard Road, Suite 607, Greenwood Village, CO 80111 (800)527-2667 

Circle 104 on reader service card 


Sleep Testing 

Imagine the people you can reach, 
the needs you can fulfill. 

Portable Apnea Recorder ^ 

CNS.Inc. 1250 Park Road, Chanhassen, MN 55317-9260. (800)843-2978. InMN (612)474-7600 In Europe: Erich Jaeger GmbH Germany, 0931/41902-0 

Circle 117 on reader service card 


hypoxemia during the procedure in 
infants. Twenty-one infants (one to 
15 weeks of age) undergoing LP for 
evaluation of possible sepsis were 
randomly assigned to the control 
group (12) or treatment group (9). 
The treatment group was preoxy- 
genated breathing oxygen (Fio: = 
1.0) spontaneously via snug face 
mask for 3 minutes prior to being po- 
sitioned for the LP. The control 
group spontaneously breathed room 
air during this interval. Oxyhemo- 
globin saturation was measured prior 
to, and continuously during, the LP 
with pulse oximetry. The groups 
were comparable in age, resting res- 
piratory rate, baseline saturation, and 
duration of the procedure. The treat- 
ment group developed significantly 
less desaturation during the pro- 
cedure than the control group (p < 
0.05). We conclude that preoxy- 
genation prior to LP prevents most of 
the hypoxemia resulting from the 
procedure in infants. 

Compliance of the Respiratory 
System in Newborn Infants Pre- 
and Postsurfactant Replacement 
Therapy— E Kelly. H Bryan, F 
Possmayer, H Frndova, C Bryan. Pe- 
diatr Pulmonol 1993;15:225. 

Surfactant administration causes a 
rapid and dramatic improvement in 
gas exchange, but, paradoxically, 
studies have failed to show an im- 
provement in the mechanical prop- 
erties of the lung. We have measured 
dynamic and static (passive flow- 
volume technique) compliance be- 
fore and after a single dose of bovine 
lipid extract surfactant in 22 pre- 
mature infants with RDS. This had 
no effect on the measured dynamic 
compliance. In contrast, surfactant 
significantly increased static com- 
pliance from 0.41 ± 0.02 to 0.55 ± 
0.04 mL/cm FLO/kg. This improve- 
ment was the result of a substantial 
recruitment of lung volume after sur- 

factant administration. This led us to 
reduce ventilator pressures, which 
produced an increase in both dynam- 
ic and static compliance, but did not 
recruit additional volume. We con- 
clude that surfactant causes a sub- 
stantial increase in static compliance 
due to volume recruitment, which is 
consistent with reports of increase in 
the measured FRC. However, despite 
this improvement, the compliance is 
still below our normal range. 

Transport Stabilization Times for 
Neonatal and Pediatric Patients 
Prior to Interfacility Transfer — 

JM Whitfield, MK Buser. Pediatric 
EmergCare 1993;2:69. 

The stabilization times for 2,863 ne- 
onatal and pediatric interfacility 
transports are reported. Appropriate 
stabilization of the sick neonate and 
pediatric patients prior to transfer is 
considered essential to reduce the ad- 
verse events that may otherwise oc- 
cur during the transfer process. The 
median stabilization time for a neo- 
nate was 80 minutes, and it was 45 
minutes for a pediatric patient. These 
times can be used by other transport 
systems to evaluate their services 
through quality assurance, as well as 
for staff and budgeting purposes. 
New interventions that reduce sta- 
bilization times can be evaluated 
with these times as a reference. 

Assessment of Forced Expiratory 
Volume: Time Parameters in De- 
tecting Histamine-Induced Bron- 
choconstriction in Wheezy Infants 

— DJ Turner, PD Sly, PN LeSouef. 
Pediatr Pulmonol 1993;15:220. 

A new technique recently introduced 
allows the measurement of infant 
lung function from lung volumes 
raised by a pump prior to generation 
of forced expiratory flow- volume 
(FEFV) curves. Forced expiratory 

volume-time (FEV,) parameters rath- 
er than traditional flow parameters 
are used. The aim of this study was 
to assess the usefulness of FE\\ pa- 
rameters in detecting airway re- 
sponsiveness to histamine in re- 
currently wheezy infants. Ten infants 
(age 7-18 months) sedated with 
80 mg/kg chloral hydrate underwent 
a histamine inhalation challenge. 
Throughout the challenge FEFV 
curves were generated from end-tidal 
inspiration. Raised FEFV curves, 
from a lung volume preset by 15 cm 
FLO inflation pressure, were gath- 
ered at baseline and after the final 
concentration of histamine. The 
mean fall from baseline was 47% in 
maximal flow at functional residual 
capacity (V maxF Rc) (p < 0.0005), 
15.5% in FEV0.5 (p < 0.0001 ), 13.5% 
in FEV0.75 (p < 0.005), and 1 1.0% in 
FEV i.o (p = 0.057), after the final 
concentration of histamine delivered. 
Tidal volume and inspiratory volume 
reached above FRC between pre- 
and posthistamine did not change. 
Mean oxygen saturation fell sig- 
nificantly from 97 to 93%. We con- 
clude that FEVt parameters adequate- 
ly detect reduced lung function 
during histamine-induced broncho- 
constriction and appear suitable for 
histamine challenge testing in in- 

Poor Agreement between Report- 
ed and Recorded Nocturnal Cough 
in Asthma — A Falconer. C Oldman. 
P Helms. Pediatr Pulmonol 1993:15: 

Reported presence or absence of 
night cough was compared with tape 
recorded cough in 15 children with 
perennial asthma (median age, 9 
years; range. 7-14) who reported 
troublesome nocturnal symptoms. 
Measurements were made and diar- 
ies kept for 7 consecutive nights. 
Cough was reported on 66 of 105 
(66%) and recorded on 93 (90%) 




available nights with poor overall 
agreement (Cohen's coefficient of 
assessment, kappa +0.30. range -0.17 
to +1). The poor agreement between 
subjective and objective assessment 
of an important symptom of noc- 
turnal asthma raises questions on the 
validity of symptom reporting and 
may in part explain the not in- 
frequent disagreement between med- 
ical and patient assessment of dis- 
ease severity. 

Accuracy of the Death Certificate 
in a Population-Based Study of 
Asthmatic Patients — LW Hunt Jr, 
MD Silverstein. CE Reed, EJ 
O'Connell, WM O'Fallon, JW Yung- 
inger. JAMA 1993;269:1947. 

OBJECTIVE: To quantify the reli- 
ability of death certificate data con- 
cerning asthma. DESIGN: The com- 
plete medical records of decedents 
were reviewed by a physician certi- 
fied in allergy and pulmonology who 
determined the cause of death with- 
out having access to the original 
death certificate. Disagreements be- 
tween the death certificate and the re- 
viewer were adjudicated by an expert 
panel. SETTING: The community of 
Rochester MN. PATIENTS: The 
mortality cohort included 339 deaths 
from a larger cohort of 5,241 Roch- 
ester residents who received medical 
treatment for asthma between 1964 
and 1983. MAIN OUTCOME MEA- 
SURES: Kappa coefficients were 
used to measure agreement beyond 
that expected by chance between the 
reviewer and the death certificate. 
The sensitivity and specificity of the 
death certificate diagnosis of asthma 
were calculated against the standard 
of the reviewer/panel diagnosis. RE- 
SULTS: Death certificates reported 
asthma as an immediate or under- 
lying cause of death in 22 instances 
(6%), whereas the reviewer/panel 
identified asthma in 53 cases (16%). 
In four cases, the death certificate 
listed asthma and the panel identified 

another cause of death. The death 
certificate had a sensitivity of 42% 
and a specificity of 99% compared 
with the reviewer/panel. Agreement 
between death certificates and the re- 
viewer was not influenced by wheth- 
er an autopsy was performed. CON- 
CLUSIONS: Death certificate diag- 
nosis of asthma as an underlying 
cause of death had a low sensitivity 
but a high specificity. Increases in 
mortality due to asthma are not likely 
caused by false-positive diagnoses of 
asthma as an underlying cause of 
death. Asthma mortality rates, de- 
termined from death certificate data, 
may indeed underestimate actual 
asthma-related mortality. 

Serum Antioxidants as Predictors 
of Adult Respiratory Distress Syn- 
drome in Patients with Sepsis — JA 

Leff, PE Parsons, CE Day, N Ta- 
niguchi, M Jochum, H Fritz, et al. 
Lancet 1993;341:777. 

Adult respiratory distress syndrome 
(ARDS) can develop as a complica- 
tion of various disorders, including 
sepsis, but it has not been possible to 
identify which of the patients at risk 
will develop this serious disorder. 
We have investigated the ability of 
six markers, measured sequentially 
in blood, to predict development of 
ARDS in 26 patients with sepsis. At 
the initial diagnosis of sepsis (6-24 h 
before the development of ARDS), 
serum manganese superoxide dis- 
mutase concentration and catalase 
activity were higher in the 6 patients 
who subsequently developed ARDS 
than in 20 patients who did not de- 
velop ARDS. These changes in anti- 
oxidant enzymes predicted the de- 
velopment of ARDS in septic pa- 
tients with the same sensitivity, spec- 
ificity, and efficiency as simultane- 
ous assessments of serum lactate de- 
hydrogenase activity and factor VIII 
concentration. By contrast, serum 

glutathione peroxidase activity and 
cti Pi-elastase complex concentration 
did not differ at the initial diagnosis 
of sepsis between patients who did 
and did not subsequently develop 
ARDS, and were not as effective in 
predicting the development of ARDS. 
Measurement of manganese super- 
oxide dismutase and catalase, in ad- 
dition to the other markers, should 
facilitate identification of patients at 
highest risk of ARDS and allow 
prospective treatment. 

Gastroduodenal Dysfunction and 
Bacterial Colonisation of the Ven- 
tilated Lung— TJJ Inglis, MJ Sher- 
ratt, LJ Sproat, JS Gibson, PM 
Hawkey. Lancet 1993;341:91 1. 

The source of ventilator-associated 
pneumonia (gastric or oropharyngeal 
flora) remains controversial. We in- 
vestigated the source of bacterial co- 
lonisation of the ventilated lung in 
100 consecutive intensive-care pa- 
tients. Gram-negative bacilli were 
isolated from the lower respiratory 
tract in 19 patients. Bacteria isolated 
from the stomach contents either pre- 
viously or at the same time were 
identical to lower respiratory isolates 
in 1 1 patients. No gram-negative 
oropharyngeal isolate was identical 
to a lower respiratory tract isolate. 
Gastric bacterial overgrowth with 
gram-negative bacilli was associated 
with the presence of bilirubin in the 
stomach contents. Detectable bili- 
rubin was also associated with sub- 
sequent acquisition of gram-negative 
bacilli in the lower respiratory tract. 
Only 5 gastric aspirate specimens 
with pH < 3.5 contained gram- 
negative bacilli. These results es- 
tablish a relation between duodenal 
reflux and subsequent bacterial co- 
lonisation of the lower respiratory 
tract. Restoration of normal gas- 
troduodenal motility might help pre- 
vent pneumonia in intensive-care pa- 



Ohmeda and oximetry. . . 

a combination you can count on, 


t Ohmeda, 
we realize how demanding 
the medical environment can be. 
And how choosing the equipment 
that best meets your needs can be 
challenging as well. 

That's why we manufacture a 
complete family of pulse oximetry 
products that offer more of what you 
expect. Like superior accuracy, easy-to- 
use features, and most importantly, reliability. 

What's more, when you purchase Ohmeda 
monitors, you're backed by an established 
market leader with a long history of technical 
innovation and excellent customer support. 
Combined with one of the largest sales and 
service organizations in the business and a 
wide variety of education and training 
materials, you can be sure that the support 
you need is always close at hand. 

When it comes to oximetry, count on the 
company you've known for years. 

Ohmeda Bio* 

Portable SpC>2 monitor with 
built-in printer for hard-copy 


1315 West Century Drive 
Louisville CO 80027 USA 
1 800 345 2700 

Circle 103 on reader service card 


Comparison of Mask and Nasal 
Continuous Positive Airway Pres- 
sure after Extubation and Me- 
chanical Ventilation — C Putensen, 
C Hormann, M Baum, W Lingnau. 
Crit Care Med 1993;2 1:357. 

OBJECTIVE: To examine the effects 
of continuous positive airway pres- 
sure applied via face masks and nose 
masks on the change in functional re- 
sidual capacity and gas exchange. 
DESIGN: Descriptive and prospec- 
tive study. SETTING: Intensive care 
unit of a university hospital. PA- 
TIENTS: Ten patients with acute 
lung injury who had required me- 
chanical ventilation. INTERVEN- 
TIONS: Continuous positive airway 
pressure at a level of 10 cm H 2 ap- 
plied in random order via face and 

nose masks. MEASUREMENTS & 
MAIN RESULTS: Both continuous 
positive airway pressure methods re- 
sulted in an almost identical increase 
of functional residual capacity. Dur- 
ing nasal continuous positive airway 
pressure, the increase in functional 
residual capacity was 294 ± 82 mL. 
During mask continuous positive air- 
way pressure, the increase in func- 
tional residual capacity was 290 ± 85 
mL. P a o; increased and the alveolar- 
arterial oxygen tension/alveolar oxy- 
gen tension quotient decreased sig- 
nificantly during mask continuous 
positive airway pressure and nasal 
continuous positive airway pressure 
at a level of 10 cm H : 0. Two pa- 
tients showed a periodic change in 
their breathing patterns; they took a 
few breaths at an increased lung 

volume, followed by one deep ex- 
piration caused by mouth opening. 
Change in mask pressure was neg- 
ligible in these two patients. Using a 
visual analog scale (10 = highly 
comfortable; = severely uncom- 
fortable), the patients rated nasal 
continuous positive airway pressure 
(8.6 ± 0.9) significantly more com- 
fortable than mask continuous pos- 
itive airway pressure (2.6 ± 0.8). 
CONCLUSION: The major advan- 
tages of continuous positive airway 
pressure (the improvement of func- 
tional residual capacity and oxygen 
transfer) can also be achieved with 
nasal continuous positive airway 
pressure in the postextubation period 
in patients who have been mechan- 
ically ventilated for acute lung in- 

Ideally You'd Like Shorter Treatment Times, 
Deeper Medication Penetration. 

>> Introducing 
yr the Ideal 



After years of research, PMP has designed 
a disposable updraft nebulizer that is 
exceptional in design and function. The 
basic nebulizer is available in five different 
configurations, including one specifically 
for environmentally-controlled therapies. 

For a FREE sample of the BetaMist2 
Nebulizer, please call us direct at 


Greenwood, SC 29646 

Circle 134 on reader service card 

BetaMist, Treatment Time Is Significantly Reduce 

■ National Brand 

■ BetaMist MHM 

Albuterol Dosage 

Small Volume Nebulizer 
Total Particle Density 

National Brand 

1150 ^H 

moo — ^H — j^H _ 

8D, 10A 80, 20A 8D, 30A 8D, 40A 8D, 50A 
FlowRatesLPM(D = Drive, A = Assist) 
Higher Particle Density Increases Potential Medication Delivery 
to the Beta Receptors in the Lungs 

At a leading university research center, 
BetaMist2™* and a leading nebulizer wen 
tested with patients suffering from mild 

©1993, Professional Medical Products, Inc. RC 692-5/ 

BetaMist2 is a trademark of Professional Medical Products. Inc. 
■Paf. Pend. 

Quality By Design 

Blood Gas/ Hemoglobin Analyzers 

AVL's uncompromising commitment is to provide the high- 
est quality Blood Gas Analyzers and CO-Oximeters through 
state-of-the-art technology and unsurpassed reliability 

The full line of AVL instruments is designed to satisfy 
every laboratory need. The AVL model 995, and the eco- 
nomical 990, for pH, PO2 and PCO2. AVL 995 Hb combines 
all the AVL 995 capabilities plus Total Hemoglobin. The 
AVL 912 CO-Oxylite measures hemoglobin fractions and 
reports calculated parameters. 

Of course, the AVL 995 and 995 Hb interface with the 
complete line of AVL Electrolyte Analyzers to provide fully 
integrated BGA and electrolyte reports. 

Advanced technology is designed into each analyzer 
to deliver the lowest annual operating cost and allow full 
testing of samples as small as 25 UL. 

The most comprehensive service coverage available 
is provided through the AVL ADVANTAGE PROGRAM: 

• The onlyfive year, full-service warranty 

• 24 hour-a-day telephone technical support 

• 24 hour-a-day on-line, computer aided instrument diag- 
nostics via a modem. If s called TeleLink. 

• Coast-to-coast field technical service representatives. 

AVL analyzers are the preferred choice of hospitals, 
clinics and laboratories worldwide. For more information 
about the complete line of quality analyzers and the exclu- 
sive Jive year, full-service warranty, write AVL Scientific 
Corporation, P.O. Box 337, 33 Mansell 
Court, Roswell, GA 30077, or call toll-free 
(800) 421-4646. In Georgia, (404) 


Circle 143 on reader service card 


The Acute Physiology and Chronic 
Health Evaluation II Classification 
System Is a Valid Marker for 
Physiologic Stress in the Critically 
111 Patient— PE Brown, SA 
McClave, NW Hoy, AF Short, LK 
Sexton, KL Meyer. Crit Care Med 

OBJECTIVE: To compare the Acute 
Physiology and Chronic Health Eval- 
uation (APACHE II) score with rest- 
ing energy expenditure obtained from 
indirect calorimetry to determine 
whether the APACHE II scoring sys- 
tem is an accurate, objective measure 
of the degree of critical illness and 
physiologic stress between groups of 
patients. DESIGN: Prospective study. 
SETTING: University hospital, ter- 
tiary referral center. PATIENTS: 
Seventy critically ill patients, con- 
secutively sampled from burn, surgi- 
cal, and medical intensive care units. 
INTERVENTIONS: Indirect calori- 
metric studies were performed on 
each patient using a metabolic cart. 
The acute physiologic score com- 
ponent of the APACHE II scoring 
system was determined at the time of 
metabolic testing, a mean of 15.9 
days after hospital admission. 
SULTS: True resting energy expend- 
iture was calculated by adjusting the 
measured energy expenditure for 
diet-induced thermogenesis and fe- 
ver. A predicted resting energy ex- 
penditure was calculated for each pa- 
tient using the Harris-Benedict 
equation alone, and by using the Har- 
ris-Benedict value corrected for pre- 
viously published metabolic activity 
factors. To eliminate differences in 
body composition and size, true rest- 
ing energy expenditure was divided 
by weight, body surface area, and 
Harris-Benedict resting energy ex- 
penditure. Results showed no signif- 
icant correlation between APACHE 
II scores and either the Harris- 
Benedict resting energy expenditure 
or the Harris-Benedict value cor- 

rected by metabolic activity factors. 
However, there was a significant (p 
0.001; r = 0.18 to 0.20) relation- 
ship between increasing APACHE II 
scores and both increasing measured 
and true resting energy expenditure. 
The true resting energy expenditure 
divided by body surface area, kilo- 
gram body weight, and Harris- 
Benedict predicted value, were all 
shown to be significantly (p < 0.01) 
related to APACHE II score, but 
showed no better degree of correla- 
tion (r = 0.12 to 0.23) than com- 
parison of APACHE II score with 
measured or true resting energy ex- 
penditure. CONCLUSIONS: The 
APACHE II classification may be a 
valid marker of physiologic stress as 
demonstrated by its statistically sig- 
nificant (although weak) relationship 
with indirect calorimetric measures 
of energy expenditure associated 
with varying degrees of critical ill- 

Early Prediction of Individual 
Outcome after Cardiopulmonary 
Resuscitation — C Madl, G Grimm, 
L Kramer, W Yeganehfar, F Sterz, B 
Schneider, et al. Lancet 1993;341: 

Prediction of individual outcome af- 
ter cardiopulmonary resuscitation is 
of major medical, ethical, and socio- 
economic interest but uncertain. We 
studied the early predictive potency 
of evoked potential recording after 
cardiac arrest in 66 resuscitated pa- 
tients who returned to spontaneous 
circulation but were unconscious 
and mechanically ventilated. Detailed 
long-latency and short-latency sen- 
sory evoked potentials were recorded 
and neurological evaluations were 
done 4-48 h after admission to in- 
tensive care. In all 17 patients with 
favourable outcome (cerebral per- 
formance categories 1 and 2) the cor- 
tical evoked potential N70 peak, a re- 
liable measure of cortical function, 
was detected between 74 and 1 16 

ms. In 49 patients with bad outcome 
(categories 4 and 5) the N70 peak 
was absent in 35 or found with a de- 
lay between 121 and 171 ms in 14 (p 
< 0.05 vs favourable outcome). Thus 
the predictive ability was 1007r with 
cutoff of 118 ms. To confirm re- 
producibility and validity, repeated 
tracings and linked-earlobe refer- 
enced techniques were done and gave 
similar results. Early recording of 
long-latency evoked potentials after 
cardiopulmonary resuscitation is 
highly predictive of outcome. 





Advertising; Dollars 


Your Association 

and Tour Profession. 


Reader Service Card 

for Information on 







This 14 hour/ 5 lb. 
OXYLITE system, 
PURSE, is only one of 
four different 
OXYLITE systems 
available to meet 
varying patient needs. 

This 2 L/M 

home oxygen patient 

is enjoying a new freedom 

to move about, travel, and lead 

a more normal life thanks to her 1 4 hour/ 

5 lb. OXYLITE complete portable oxygen 


The heart of the 

OXYLITE system is 


electronic oxygen 

conserver which 

significantly extends 

the duration of the 

small cylinder 

provided with the 

OXYLITE system. 

For more information on OXYLITE 
Systems, the Superior Alternative to 
Liquid Oxygen, Call: 



9445 DeSoto Ave. 
Chatsworth.CA 91311 
ERAPBJTICS, INC. (818)882-0883/(800)423-8870 

Circle 109 on reader service card 



.* V'i 


»■« * 

4 -*\Jl J 

Group Photo (left to right) 


David J Pierson 


Frank B Cerra 

Back row: David J Pierson, Pat Brougher, Sam Giordano, Donna 

Leonard D Hudson 

Charles G Durbin Jr 

Stephens, Loren D Nelson, James K Stoller, Ray Masterrer, Leonard D 


Ray Masferrer 

YC Tony Huang 

Hudson. David R Dantzker, Doug Mclntyre, Alan H Morris, and Thomas 

Thomas D East 

Yui-Ling Tang 

D East. 


Alan M Morris 

Loren D Nelson 

David R Dantzker 


P Terry Phang 

Front row: Lawrence DH Wood, Frank B Cerra, Dean Hess, P Terry 


Loren D Nelson 

Thomas D East 

Phang, Richard D Branson, Charles G Durbin Jr, Paul J Mathews, Jon 

James K Stoller 


Lawrence DH Wood 

Nilsestuen, and Larry H Conway. 


Richard D Branson 


Dean Hess 


Frank B Cerra 

Jon Nilsestuen 


David R Dantzker 


Charles G Durbin Jr 



A Special Issue 

based on a conference 
held October 8- 10. 1992. 
in Puerto Vallarta, Mexico 

sponsored by 

The American Association for Respiratory Care 

and its science journal RESPIRATORY CARE 

with the aid of an education grant 
from Allen & Hanburys, Div of Glaxo Inc 

Chairmen and 
Guest Editors: 


Dean Hess MEd RRT and David J Pierson MD 

Richard D Branson RRT 
Frank B Cerra MD 
David R Dantzker MD 
Charles G Durbin Jr MD 
Thomas D East PhD 
Dean Hess MEd RRT 
Thomas L Higgins MD 

Leonard D Hudson MD 
Alan H Morris MD 
Loren D Nelson MD 
P Terry Phang MD 
David J Pierson MD 
James K Stoller MD 
Lawrence DH Wood MD 

Discussants: The chairmen, the speakers, YC Tony Huang MD, Larry H Conway RRT. 

and Paul J Mathews MPA RRT 


Dear Reader: 


1 1030 Abies Lane, Dallas, TX 75229 • 214/243-2272, Fax 214/484-2720 

The American Association for Respiratory Care and its science journal 
Respiratory Care are proud to present the June and July 1993 issues 
containing the proceedings from the 1992 Journal Conference— 
Oxygenation in the Critically 111 Patient. The papers and discussions 
presented here are well-organized, state-of-the-art materials designed 
for you to use in your daily practice. 

Considerable planning and work are required to present and publish the 
proceedings of a conference like this. We are grateful to the 
speakers, the Editorial Board, Conference Co-Chairmen Dean Hess MEd 
RRT and David J Pierson MD, Editor Pat Brougher, and Editorial 
Coordinator Donna Stephens for planning, writing, editing, and 
participating in the Conference and sharing this important information 
with us. 

We are indebted to Allen & Hanburys, Division of Glaxo Inc, for their 
generous grant to the Association, which made possible the 
presentation and publication of the Conference proceedings. Allen & 
Hanburys has supported the last five Journal Conferences, for which we 
are enormously grateful . 


Ray Masferrer RRT 
Managing Editor 
Respiratory Care 

Conference Proceedings 

Normal and Abnormal Oxygenation: 
Physiology and Clinical Syndromes 

David J Pierson MD 








Acute Respiratory Failure as a Problem of 
Abnormal Arterial Blood Gas Values 
Clinical Causes of Hypoxemia 

A. Alveolar Hypoventilation 

B. Ventilation-Perfusion Mismatch 

C. Right-to-Left Shunt 

D. Hypoxemia without Acute Respiratory Failure 
Respiration as a Tissue Process 

Acute Respiratory Failure as a Problem of 

Global Tissue Oxygenation 

Clinical Categories of Oxygenation Failure 

A. Decreased Blood Oxygen Content 

B. Decreased Cardiac Output 

C. Impaired Tissue Oxygen Utilization 
Tissue Oxygenation — Insights and Problems 
Acute Respiratory Failure as a Problem of 
Local Tissue Oxygenation 


The outcome of critical illness depends more 
on the adequacy of oxygenation than on any other 
factor. No aspect of respiratory care commands 
more of the clinician's attention or expends more 
resources. Abnormal oxygenation is central to the 
concept of acute respiratory failure, and all ap- 
proaches to management, however disparate, focus 
on its correction. 

Dr Pierson is Medical Director, Respiratory Care. Harborview 
Medical Center, and Professor of Medicine, University of 
Washington — Seattle, Washington. 

A version of this paper was presented by Dr Pierson on Oc- 
tober 8. 1992, during the Respiratory Care Journal Confer- 
ence on Oxygenation in the Critically 111 Patient, held in Puerto 
Vallarta. Mexico. 

Reprints: David J Pierson MD, Harborview Medical Ctr, ZA- 
62, 325 Ninth Avenue, Seattle WA 98104. 

Both our understanding of impaired oxygenation 
in patients and the methods for applying this under- 
standing at the bedside have evolved substantially 
during the last quarter century. In this paper, I trace 
this evolution to clarify current concepts of acute 
oxygenation failure and emerging approaches to as- 
sessment and management. I first review the ele- 
ments of respiration and the definition of acute res- 
piratory failure that emerged following the clinical 
introduction of arterial blood gas (ABG) analysis. 
After describing the mechanisms of hypoxemia and 
illustrating these with brief clinical examples, I dis- 
cuss the evolution from ABG values to global tis- 
sue oxygenation that has dominated thinking in this 
field during the last decade. 

Using further examples of acute respiratory fail- 
ure that are less well explained by the traditional 
ABG model, I emphasize the importance of addi- 
tional mechanisms for impaired tissue oxygenation. 
Finally, in light of emerging techniques for assess- 




ing regional oxygenation, I touch on some of the 
new issues faced by clinicians in defining, evaluat- 
ing, and managing patients with acute respiratory 

Acute Respiratory Failure as a Problem of 
Abnormal Arterial Blood Gas Values 

Clinicians describe disease and devise therapies 
with the tools at hand, and in medicine these tools 
are continually changing. With each introduction of 
technology, come new definitions, diagnostic cri- 
teria, and treatments for the same conditions. Prior 
to the introduction of ABG analysis, clinical assess- 
ment was descriptive and largely external. The 
present-day concept of acute respiratory failure as a 
clinical syndrome of life-threatening derangement 
of respiratory function, something that could de- 
velop in any number of different disease settings, 
was unknown. 1 The severity of lobar pneumonia or 
other specific illness was described in terms of 
symptoms and signs, plus various indirect assess- 
ments of respiratory function such as chest radio- 
graphs and blood leukocyte counts. 

Clinical introduction of ABG analysis in the 
1960s permitted more direct assessment of respira- 
tory function, and a physiologic definition for acute 
respiratory failure emerged based on the lungs' role 
in blood oxygenation and carbon dioxide (C0 2 ) re- 
moval. 2 In 1965, EJM Campbell wrote 1 

The function of the respiratory system is to secure gas 
exchange between blood and ambient air so that the ar- 
terial blood gas pressures are kept within certain limits. 
Therefore respiratory failure may be defined as impair- 
ment of this function to such degree that the arterial 
blood gas pressures depart from these limits. 

In the 1970s, ABG-based definitions of acute 
respiratory failure were widely used, although 
those of different authorities varied in their thresh- 
old values for arterial partial pressure for oxygen 
(P a o 2 ) and for C0 2 (P a co 2 )- A 1971 statement is 
shown below. 1 

Arbitrarily, acute respiratory failure is defined as the 
sudden development of an arterial Po? of less than 50 
mm Hg. with or without CO: retention. 

And, in 1977, Murray said the following: 4 

Acute respiratory failure is usually defined on the basis 
of alterations in arterial blood gas composition: an ar- 

terial Po; < 60 mm Hg and/or an arterial Pco: > 50 mm 

For acute respiratory failure defined by hypox- 
emia (a PaO: below the normal range), although the 
defining P a o2 values recommended by different au- 
thorities varied, 5 the concept emerged of a clin- 
ically significant fall in arterial oxygenation. How 
much of a decrease in P a o: was clinically important 
was dictated by the oxygen-hemoglobin dis- 
sociation curve (Fig. I). 6 Once acute illness had 
caused P a o: to fall below about 60 torr, the at- 
tendant decrease in arterial hemoglobin saturation 
(S a cb) implied a potentially life-threatening fall in 
blood oxygen content. ftJ Similarly, the diagnosis of 
acute ventilatory failure rested on a rise in P a co 2 
that was rapid enough and of significant magnitude 
to produce a clinically important fall in pH, typ- 
ically to below 7.30. 78 

/ — 


Total________- — ■•' _ - 

S*^ Carried by Hb 



3? 6 °- 




(fi 40- 





y Dissolved 

12 f 

20 40 60 80 100 600 

Pa0 2 (torr) 

Fig. 1. Relationship among arterial P02 (Pa02, horizontal 
axis), oxyhemoglobin saturation (Sa02, left vertical axis), 
and arterial 2 content (C a 02. right vertical axis). As Pa02 
falls, both Sa02 and C a 02 diminish rapidly once the 'shoul- 
der' of the curve is reached, at Pa02 values below 50 to 
60 torr. This phenomenon is the basis for defining hypox- 
emic acute respiratory failure in terms of a Pa02 value in 
this range or below. (Reproduced from Reference 6, with 

Once ABGs became widely used to assess the 
respiratory status of acutely ill patients, acute res- 
piratory failure was divided conceptually into two 
basic categories. These were, and remain, failure of 
oxygenation manifested by hypoxemia and failure 
of ventilation manifested by acute respiratory ac- 




idosis. The rest of this article focuses on the for- 
mer; comprehensive reviews of acute ventilatory 
failure are readily available elsewhere. 9 " 12 

Clinical Causes of Hypoxemia 

Five physiologic mechanisms can result in hy- 
poxemia (Table 1): a decrease in inspired P02 
(P102), diffusion limitation, alveolar hypo- 
ventilation, ventilation-perfusion (Va/Q) mismatch, 
and right-to-left shunt. Low Pio: is an uncommon 
cause for acute respiratory failure but may be en- 
countered in fires (due to consumption of ambient 
: ), in acute high altitude exposure, or in the in- 
tensive care unit (ICU) when supplemental : ther- 
apy is inadvertently or unwisely interrupted. Al- 
though it can be a cause for hypoxemia during 
exercise at extreme altitude, diffusion limitation is 
not a clinical cause of acute respiratory failure as 
encountered in the ICU. For clinical purposes, hy- 
poxemia is caused by one or more of the three re- 
maining mechanisms — alveolar hypoventilation, 
Va/Q mismatch, and right-to-left shunt. Distin- 
guishing among these is both clinically important 
and readily accomplished. 

Table 1. Hypoxemic Respiratory Failure: Physiologic Mech- 
anisms and Clinical Settings 


Clinical Examples 

Low inspired Prj; 

High altitude 

Patient not receiving prescribed O: 

Alveolar hypoventilation Narcotic overdose 

Acute COPD exacerbation 


Right-to-left shunt 

Acute COPD exacerbation; asthma 
Diffuse lung disease 

Generalized process 


Cardiogenic pulmonary edema 
Localized process 

Lobar pneumonia 


Alveolar Hypoventilation 

In the presence of a normal alveolar-to-arterial 
P02 difference [PiA-aiO:]. alveolar hypoventilation 
produces a fall in P a o: that is roughly equivalent to 
the increase in P a co:- Because alveolar and arterial 

CO : tensions are essentially identical, P a co: is used 
to determine the presence and magnitude of alveo- 
lar hypoventilation (Figs. 2 & 3). 6,13 

Fig. 2. Schematic depiction of how Pa02 and PaC02 
change in opposite directions, assuming an unchanging 
P(A-a)02 and a respiratory exchange ratio (C0 2 produced/ 
O2 consumed) of 0.8. Alveolar hypoventilation raises 
Paco2 above the normal value of 40 torr and diminishes 
Pa02 proportionally. Hyperventilation (dashed line) pro- 
duces the opposite effect. (Adapted from Reference 6, 
with permission). 

Case 1: Hypoxemia due to alveolar hypoventila- 
tion — A young man with a history of previous suicide 
attempts is found unresponsive with several empty pill 
bottles nearby. He is cyanotic and taking only shallow 
breaths 8 times per minute. ABG values while breathing 
room air (RA) show Po: 46 torr, Pco: 72 torr. and pH 
7 22 

Comment: The Ra-uiO: is approximately 14 torr (nor- 
mal), indicating that this patient's hypoxemia can be ac- 
counted for entirely by alveolar hypoventilation. 

Often one or both of the other mechanisms are 
present in addition to alveolar hypoventilation, as is 
typically seen in chronic obstructive pulmonary 
disease (COPD). 13 In such instances, the hypox- 
emia is multifactorial in pathophysiology, with al- 
veolar hypoventilation accounting for a reduction 
in P a o: in proportion to the elevation in P a co: above 
40 torr. 

Ventilation-Perfusion Mismatch 

Although the matching of alveolar ventilation to 
pulmonary perfusion normally varies somewhat in 
different regions of the lung, the overall effect is 
one of approximately equal matching, although W 





Fig. 3. Conceptual diagram of hypoxemia due to alveolar 
hypoventilation. Both idealized lung units depict oxy- 
genation on the left and ventilation (C0 2 exchange) on 
the right. In this example P(A-a)02 is assumed to be zero. 
A. Normal gas exchange. B. Aveolar hypoventilation, with 
a reduction in alveolar ventilation to one half normal. (Re- 
produced from Reference 6, with permission.) 

Q is somewhat less than 1 (approximately 0.9). 
When overall ventilation is diminished in pro- 
portion to perfusion (Va/Q < 1), hypoxemia results, 
as shown in Figure 4A. 6 Areas of low Va/Q have 
some ventilation (otherwise they would be right-to- 
left shunt areas), but not enough to fully saturate 
the hemoglobin in the blood passing through them. 
Mismatching of Va/Q is the most common cause of 
hypoxemia, in both clinically stable and acutely ill 
individuals. Even in patients with COPD and al- 
veolar hypoventilation, VA/Qmismatch generally ac- 
counts for most of the decrease in P a 02- 

Hypoxemia due to Va/Q mismatch tends to be 
relatively easy to correct, as shown schematically 
in Figure 4B. 6 By increasing the "driving pressure" 
of 2 into the poorly ventilated alveoli, one can 
raise P a cb sufficiently to fully saturate the venous 
blood in the capillaries. This is why patients with 
COPD or acute asthma can usually be managed 
with only modest amounts of supplemental 2 (see 
Case 2), in contrast to patients with hypoxemia due 
to shunt. 

V/Q normal 

V/Q low 

P0 2 = 40 
2 Content = 15 

PO 2 =40 

O, Contents 15 

Fig. 4. Hypoxemia due to WQ mismatch, illustrating the 
effect of supplemental 2 on Pa02- WQ is normal on the 
left side of each idealized lung unit, and low on the right. 
Only O2 exchange is shown, and P<A-a)02 is assumed to 
be zero. A. During breathing of room air (Pio 2 = 150 torr), 
not enough 2 reaches the poorly ventilated alveolus to 
fully saturate its capillary blood. B. With 40% O2 (P102 = 
285 torr), alveolar P02 is raised sufficiently to make the 
capillary P02 nearly normal. Pa02 as measured by ABGs 
reflects the 2 content of the mixed effluent from all lung 
units, as shown. (Reproduced from Reference 6, with 




Case 2: Hypoxemia due to both alveolar hypo- 
ventilation and Va/Q mismatch — An elderly woman 
with longstanding, severe COPD experiences an acute 
exacerbation, with increasing dyspnea and a change in 
sputum color over 2 days. On presentation she is in 
moderate respiratory distress with pulse 120/min, cya- 
nosis, and signs of hyperinflation. ABG values on air 
and then breathing nasal O: at 2 L/min are as shown: 






Comment: This patient has both alveolar hypo- 
ventilation and Va/Q mismatch. Her P(A-aiO: is ap- 
proximately 33 torr. indicating that her hypoxemia is in- 
itially due more to the latter than to the former. Low- 
flow supplemental O: easily corrects the hypoxemia — 
too much so, in that with a P a 02 of 88 torr, the P a co: ris- 
es by 6 torr, dropping arterial pH into a potentially dan- 
gerous range. Whether this is due to hyperoxic suppres- 
sion of ventilatory drive or further Va/Q derangement is 
unclear, 12 " 14 but it can be corrected by giving only 
enough supplemental O; to raise the P a o: into the 55-60 
torr range. 910 

Although in this example both alveolar hypo- 
ventilation and Va/Q mismatch are present, in many 
cases hypoxemia is due to the latter alone. A fa- 
miliar example is acute asthma, in which P a 02 is 
typically reduced into the 50s. whereas P a co 2 is 
commonly low-normal or below. 915 

Right-to-Left Shunt 

If blood passes through the lung without being 
exposed to any alveolar ventilation (as occurs when 
alveoli are collapsed or filled with fluid), the blood 
leaves the lung still deoxygenated. This results in 
hypoxemia through the mechanism depicted in Fig- 
ure 5A. 6 Unlike the situation with Va/Q mismatch, 
hypoxemia in the presence of a right-to-left shunt is 
not readily corrected with supplemental oxygen 
(Fig. 5B) because Pao 2 is irrelevant in lung regions 
with no ventilation at all. This observation is of 
practical clinical value, in that it enables the cli- 
nician to predict the need for high concentrations of 
supplemental 2 or other forms of therapy such as 
positive end-expiratory pressure (PEEP). 7 " 1 

Removal of C0 2 is not usually affected when 
hypoxemia is due to Va/Q mismatch, and P a c02 is 


O, Content- 15 

0,Corrtent ■ 15 

Fig. 5. Alveolocapillary diagram of intrapulmonary right- 
to-left shunt, illustrating why supplemental 2 therapy 
fails to correct the resultant hypoxemia. As in Figure 4, 
only O2 exchange is shown, and P(A-a)02 is assumed to 
be zero. A. During breathing of room air, although blood 
leaving the normal alveolocapillary unit is normally sat- 
urated, blood passing through the capillary on the right 
'sees' no 02 and remains desaturated. When the two 
streams mix, the resulting Pa02 is determined by the aver- 
age of their O2 contents, not by their P02 values. B. Addi- 
tion of 40% 2 has little impact on the hypoxemia be- 
cause O2 content is not substantially raised in the normal 
unit (its hemoglobin is already saturated) and capillary 
blood in the unventilated unit still sees no O2. Even 100% 
2 would not completely reverse the oxygenation defect 
in this example. (Reproduced from Reference 6, with per- 

usually normal or low unless there is another rea- 
son for hypoventilation. This is because of the 
shape of the C0 2 dissociation curve, which makes 
it easier for C0 2 to leave the capillary blood than 
for 2 to enter it, and also because patients typ- 
ically augment ventilation in response to hyper- 
capnia. 17 ' 18 The same is true with right-to-left shunt 
except when its magnitude is very large. 




Case 3: Hypoxemia due to right-to-left intrapulmo- 
nary shunt — A 40-year-old man with severe alco- 
holism is brought to the hospital after complaining to 
companions of fever and chills for 2 days. He is pro- 
ducing purulent, rusty sputum, and physical examina- 
tion reveals signs of marked right-upper-lobe consolida- 
tion. A chest radiograph confirms the presence of dense 
consolidation of that lobe. ABG analysis shows the fol- 





: , 2 L/nun 

: , 1007c mask 


Comment: Unlike the situation in Case 2, low-flow 
supplemental Ch has essentially no effect. Even 100% 
O: by mask fails to correct the hypoxemia, dem- 
onstrating that in this case the hypoxemia is due to 
right-to-left shunt. Administration of PEEP is not likely 
to help in this case of localized pulmonary disease, al- 
though positioning the patient with the affected lobe up- 
permost may help to decrease the perfusion (and hence 
the shunt) to that area. 1619 - 20 

Physiologic studies demonstrate that some in- 
stances of shunt as defined here, particularly in the 
adult respiratory distress syndrome (ARDS), are 
partly or entirely due to very low WQ regions. 21 * 22 
However, this is of little concern to the clinician 
managing hypoxemic patients because the re- 
sponses to therapy are the same as with 'true' right- 
to-left shunt. 16 

Hypoxemia without Acute 
Respiratory Failure 

Use of the traditional ABG definition of acute 
respiratory failure becomes problematic in some 
circumstances. Consider the following case ex- 

Case 4: Hypoxemia without acute respiratory fail- 
ure — A 49-year-old otherwise healthy man is a pas- 
senger in an airliner flying at 12,000 ft above sea level. 
Cabin pressurization. which had previously maintained 
ambient pressure the same as sea-level barometric pres- 
sure, is suddenly lost. The following changes occur in 
cabin pressure (eg, barometric pressure, Pr), PiO;. and 
arterial oxygenation, prior to any compensatory hyper- 



— > 















— > 



The passenger" s cardiac output (Q t ) promptly increases, 
which maintains systemic O: delivery (Do:) in the nor- 
mal range: 













— » 



Comment: A healthy person would also promptly hy- 
perventilate in response to the hypoxemia, raising P a o; 
accordingly (see Fig. 2). The subject in this example 
would probably be uncomfortable, but is he in acute res- 
piratory failure? His arterial Po: certainly qualities, but 
with only modest augmentation of Qt the overall deliv- 
ery of O: to his tissues is normal. 

This hypothetical example illustrates a problem 
with the traditional ABG definition of acute res- 
piratory failure and prompts reconsideration of the 
definition of respiration itself. 

Respiration as a Tissue Process 

The concept of respiration as a strictly pul- 
monary process ("In with the good air, out with the 
bad air") and the definition of acute respiratory fail- 
ure in only these terms are inadequate in the con- 
text of current understanding and ICU technology. 
A more clinically relevant definition of respiration 
is that given by Webster's International Dic- 
tionary: 23 

Respiration: the physical and chemical processes by 
which an organism supplies its cells and tissues with the 
oxygen needed for metabolism and relieves them of the 
carbon dioxide formed in energy-producing reactions. 

Figure 6 24 depicts the pathway for 2 from ambi- 
ent air to intracellular metabolism, illustrating the 
importance of considering respiration as a tissue- 
based process rather than one that occurs only in 
the lungs. 25 Figure 7 26 shows schematically the 
progressive reduction in Pcb from inspired air to 
mitochondria that occurs during the process di- 
agrammed in Figure 6. 

Acute Respiratory Failure as a Problem of 
Global Tissue Oxygenation 

In keeping with the process of respiration as just 
defined and described, acute oxygenation failure 




f-<—\i^L^ \ / V / ^- Alveolar-capillary interface 

- Pulmonary vessels 

■ Systemic vessels 
\\\ t L ■ / •> Capillary-tissue interface 

Intracellular respiratory engine'' 

Fig. 6. The pathway for 2 from outside air to ultimate 
consumption within the mitochondria of cells. Interruption 
of this pathway at any point can produce acute res- 
piratory failure at points beyond the interruption. Pa = al- 
veolar P02; Pi = inspired (eg, room air) P02; Pa = arterial 
P02; Pv = venous P02; ADP = adenosine diphosphate; 
and ATP = adenosine triphosphate. (Adapted from Refer- 
ence 24, with permission.) 

may be defined as any impairment of tissue oxy- 
genation of sufficient acuteness and severity as to 
threaten the life of the individual. Although the 
clinical tools for using it were not as readily avail- 
able as they are today, this tissue-based concept 
was described more than a quarter century ago by 
Bendixen et al, 2 who listed three types of dis- 

-*■ Mitochondria 


Fig. 7. Illustration of the progressive fall in Po 2 as 2 
passes from inspired air to alveolus, from alveolus to ar- 
terial blood, and then from arterial blood to respiring tis- 
sues. The P02 drop produced by diffusion across the al- 
veolocapillary membrane and into the erythrocyte is very 
small, as is that due to the normal ('physiologic') right-to- 
left shunt accounted for by the bronchial and thebesian 
veins. Cap = capillary; Art = artery. (Modified from Refer- 
ence 26, and reproduced with permission.) 

turbance producing hypoxic acute respiratory fail- 

( I ) Failure to maintain an adequate arterial oxygen ten- 
sion, caused by abnormal ventilation-perfusion re- 
lationships in the lung. (2) Failure to maintain an ad- 
equate arterial oxygen content, caused by a decreased 
oxygen-carrying capacity of the blood and a decreased 
arterial oxygen tension. (3) Failure to transport oxygen 
to the tissues, caused by decreased cardiac output and 
impaired distribution of blood flow. 

To this list, current classifications of oxy- 
genation failure add a fourth category, that of fail- 
ure of the peripheral tissues to take up and utilize 
the oxygen delivered to them (Table 2). 

Table 2. Clinical Forms of Oxygenation Failure 



Failure to oxygenate arterial blood 


Inadequate C a O: 
Failure of tissue oxygenation 

Inadequate Do: 

Inadequate Vo; 

iP a02 

iS a o: and/or IHb 

J-CaO: and/or IQt 

Clinical Categories of Oxygenation Failure 
Decreased Blood Oxygen Content 

This is most often due to hypoxemia, but a de- 
crease in the oxygen content of arterial blood 
(C a o:) can also occur without a reduction in P a o:- 

Case 5: Acute oxygenation failure without hypox- 
emia (carbon monoxide poisoning) — An elderly wom- 
an is pulled unresponsive from her smoke-filled apart- 
ment. She is hyperpneic and tachycardic but not 
cyanotic. Her initial ABG values breathing room air are 
Po; 90 torr, Pco; 32 torr, pH 7.25. S a O: is reported as 
96%-. A carboxyhemoglobin (COHb) level is ordered 
and subsequently reported to be 45%. 

Comment: This patient's blood hemoglobin (Hb) con- 
centration is 13.6 g/dL. so that her 'effective' Hb (the 
amount available to carry O:) is 45% less than this, or 
7.4 g/dL. Total C a o: is calculated according to the for- 

C j0: = ( l.34)(Hb)(S a 0;> + (0.003 KPaO;). [1] 

where the first product represents the O; bound to Hh 
and the second product is the amount of dissolved O;. 




1.34 is the 02-carrying capacity of Hb (in mL/g). and 
0.003 is the solubility of O: in plasma (in mL/torr). 

In this example, 

C a0: = (1. 34 K7.4 1(0.96) + (0.003X90) = 9.78 mL/dL. 

which represents a severe reduction from the 
C a O; of 17.8 mL/dL this patient would have with- 
out the COHb. For this patient to maintain a nor- 
mal Do:- she would have to compensate by sub- 
stantially increasing her cardiac output; her 
concomitant metabolic acidosis suggests that this 
is not occurring. 

In the ICU, patients frequently have reductions 
in Hb concentration following trauma, surgery, or 
other circumstance causing blood loss. This may 
result in acute respiratory failure at the tissue level 
because of decreased C a 02^ even in the absence of 
hypoxemia. Figure 8 illustrates the effects of differ- 
ent Hb concentrations on CaO:- even at high arterial 
Po2- 6 The figure shows why the use of high inspired 
2 concentrations, PEEP, or other measures to in- 
crease arterial P02 is of little benefit in such cir- 
cumstances. Once the Hb is fully saturated, which 
for practical purposes means a P a o2 > 80-90 torr, 
short of the use of hyperbaric : little increase in 
C a 02 can be achieved by raising P02 further. On the 
other hand, increasing the Hb concentration by 5 g/ 
dL by means of erythrocyte transfusion will yield a 
substantial increase in C a 02 at any P a o2- 

Hb 1 5 g/dL 

20 40 60 80 100 120 140 
Pa02 (torr) 

Fig. 8. The relationship between C a o2 and arterial P02 as 
a function of blood hemoglobin (Hb) concentration. In the 
presence of anemia, transfusion may be far more ef- 
fective in increasing C a 02 than PEEP or other measures 
to drive up arterial Po2- (Reproduced from Reference 60, 
with permission.) 

A marked reduction in C a o2 does not necessarily 
imply tissue oxygenation failure, as illustrated by 
the following example: 

Case 6: Decreased C a 02 without acute respiratory 
failure (severe anemia) — A 23-year-old woman com- 
plains of fatigue and dyspnea on exertion. She has a 6- 
month history of unusually heavy, prolonged menstrual 
periods. Examination reveals only pallor and tachy- 
cardia and is otherwise normal. Her hematocrit is 15 
volumes % and Hb 5.6 g/dL. Her P a rj2 is 90 torr, PaCO: 
40 torr, and pH 7.40. With S a o 2 98%, her C a 02 is ( 1 -34) 
(5.61(0.98) + (0.0031(90) = 7.62 mL/dL. Normal content 
(Fig. 1 ) is approximately 20 mL/dL. Is she in tissue oxy- 
genation failure? 

Comment: This patient has a marked reduction in C a O: 
because of severe anemia. Whether this represents acute 
respiratory failure depends on whether she is able to 
maintain Do:: 

Do: = CaO: x Qt. 


In this case the patient's Qt, which has increased from 
the normal level of 5.0 to 13.2 L/min. compensates for 
the reduction in C a o: to keep D02 within normal limits: 
D02 = (7.62)( 13.2) = 1.0 L/min. She is not in acute res- 
piratory failure. 

Unlike the otherwise healthy person in this ex- 
ample, critically ill patients with severe anemia are 
often unable to augment Q t sufficiently to com- 
pensate for their reduced 2 -carrying capacity. 27 
Anemia is especially deleterious in the presence of 
refractory hypoxemia, as may occur in severe 
ARDS following trauma, for example. 

Decreased Cardiac Output 

A reduction in Q, can produce tissue hypoxia 
even in the presence of normal arterial oxygena- 
tion: 28 

Case 7: Acute respiratory failure without hypoxemia 
(reduced cardiac output) — A 58-year-old man de- 
velops crushing substernal chest pain and is admitted to 
the coronary care unit with an acute anterior myocardial 
infarction. He is tachypneic and his blood pressure is 
95/60 mm Hg. He has physical signs of heart failure, 
and a chest radiograph shows pulmonary edema. While 
breathing supplemental O: at 4 L/min. the patient has a 
Pa02 of 90 torr, P a c02 of 30 torr, and pH of 7.34. The pa- 
tient's C a 02 is 20 mL/dL. which is normal. However, his 
Qt is 3.0 L/min (normal for this patient = 5.0 L/min). so 
that his Do: is (201(3.0) = 0.6 L/min. 




Comment: This patient is in acute respiratory failure, 
defined as a life-threatening reduction in tissue oxy- 
genation. His metabolic acidosis most likely represents 
lactic acidemia from inadequate tissue O; delivery. Cor- 
rection of oxygenation failure in this instance requires 
augmenting Q rather than further increasing the already 
adequate P a O: and C a 02- 29 

Oxygenation failure due to reduced Q t can be 
seen without intrinsic cardiac disease. The most 
common example of this in respiratory care occurs 
with the application of PEEP, as illustrated in Table 
3. In the example shown in the table, systemic : 
transport and hence tissue oxygenation are wors- 
ened when PEEP is applied, despite the improve- 
ment in R1O2 and C a 02- This example illustrates the 
importance of careful hemodynamic monitoring 
during the use of PEEP, especially initially and par- 
ticularly in the presence of hypovolemia. 1619 

Table 3. Acute Respiratory Failure due to Decreased Systemic 
Oxygen Transport in the Absence of Hypoxemia 

Adverse Effects of PEEP 

(cmH 2 0) 



C a O: 0, 
(mL/dL) (L/min) 






17.5 6.0 
20.2 3.5 


Impaired Tissue Oxygen Utilization 

The last form of acute respiratory failure without 
hypoxemia is less common than those just de- 
scribed but may nonetheless be encountered in the 
ICU. Several circumstances can produce a defi- 
ciency in tissue-0 : uptake and utilization despite 
normal 2 delivery to those tissues. 3 " These circum- 
stances include a marked left shift in the oxyhe- 
moglobin dissociation curve (Fig. 1), as occurs in 
severe carbon monoxide poisoning, 3132 and may 
also be seen initially after transfusion of large quan- 
tities of stored blood that has lost most of its 2,3- 
diphosphoglycerate (2,3-DPG), particularly in the 
presence of alkalemia. 30 However, the classic sce- 
nario for reduced tissue : utilization, or : uptake 
(Vo 2 ) is poisoning of the intracellular oxidative en- 
zyme system, as occurs in cyanide poisoning. 32 ' 35 

Case 8: Impaired tissue oxygen utilization (nitroprus- 
side-associated cyanide intoxication) — A 60-year-old 
woman with hypertension and chronic renal in- 
sufficiency is admitted to the ICU in hypertensive crisis 
with severe congestive heart failure. She is treated with 
a continuous intravenous infusion of sodium nitro- 
prusside. which (in order to control her blood pressure) 
must be titrated upward in dosage over 24 hours, from 
0.5 to 6.0 /Jg • min _1 ■ kg" 1 . She becomes increasingly 
restless and confused, with muscle spasms and cardiac 
dysrhythmias. Her ABG values as she breathes O; at 4 
L/min show Po: 140 torr, Pco: 26 torr. and pH 7.22. A 
mixed venous blood gas shows Po: 65 torr (normal 40). 
Pco: 32 torr, and pH 7.15. Her calculated blood-anion 
gap is markedly elevated. 

Comment: This patient's mental-status changes, car- 
diac dysrhythmias, muscle spasms, and anion-gap meta- 
bolic acidosis can all be explained by cyanide toxicity 
related to nitroprusside administration. The abnormally 
high mixed venous Po; (PvO:) is compatible with re- 
duced peripheral O; extraction due to poisoning of the 
intracellular oxidative enzyme system by cyanide. Ni- 
troprusside is metabolized to thiocyanate, which is sub- 
sequently eliminated over a half-life of 2 to 3 days. Pa- 
tients with renal insufficiency are especially susceptible 
to cyanide toxicity due to impaired excretion of the me- 

The different clinical forms of acute oxy- 
genation failure discussed here and illustrated with 
clinical examples are summarized in Table 4. 

Table 4. Comparison of Clinical Forms of Acute Oxygenation 


Blood Cardiac Tissue 

(Anemia) (CHF)* (Cyanide) 

C a O: 




heart fai 


lure; nl 

= withi 

nl nl 
nl nl 
i nl 

i I 

*CHF = 

■■ congestive 

,n normal limits. 

Tissue Oxygenation — Insights and Problems 

A related subject of intense current interest, in 
which decreased peripheral : utilization plays a 
central role, is pathologic supply-dependency of 
j^ 30.38-40 Norman^ Vo 2 is a reflection of the 
body's metabolic needs, not of Db 2 . When healthy 
tissues are subjected to a reduced : supply, they 




continue to extract 2 normally until Db 2 reaches a 
critical level. Below this critical threshold, Vo 2 be- 
comes a function of Do 2 ' and tissue damage and 
cell death may occur. In otherwise healthy in- 
dividuals this critical Db2 threshold is not reached 
until V 02 becomes 60-70% of Drj 2 . In ARDS. septic 
shock, and some other disease states, this homeo- 
static independence of Vo 2 from D02 appears to be 
altered 30 - 38 - 41 " 45 (Fig. 9). Under such circumstances 
pathologic supply-dependency of Vo 2 may be 
present when the 2 extraction ratio (Vo 2 /Eb 2 ) is 
much lower than the normal 60-70% threshold. 
Many investigators currently believe that patholog- 
ic supply-dependency of Vq 2 will prove to be a key 



Fig. 9. Conceptual depiction of the relationships between 
2 supply (D02) and 2 utilization (Vb 2 ) in both normal in- 
dividuals and critically ill patients. In the face of falling 
supply, tissues normally continue to extract 2 normally 
until this extraction becomes 60-70% of supply. In ARDS 
and perhaps in other critical illness, this ability of the tis- 
sues to regulate Vo 2 appears to be disrupted, and sup- 
ply-dependency is present over a much wider range of 
Do 2 . (Reproduced from Reference 30, with permission.) 

factor in the development of multiple systems or- 
gan failure and in the high mortality of ARDS and 
septic shock. 39 ' 41 

The mechanism or mechanisms for pathologic 
supply-dependency of Vo 2 remain unclear, and its 
very existence in critically ill patients is still de- 
bated by some. Many studies have determined both 
Vo 2 and Do 2 from some of the same measurements, 
which could accentuate or produce significant ar- 
tifactual associations among the variables, 40 ' 46,47 
and some studies that have determined Vo 2 and Do 2 
by entirely separate techniques have failed to 
demonstrate the relationship. 48 Nonetheless, this 
topic presently commands more interest in critical 

care than any other aspect of acute respiratory fail- 
ure, and it focuses attention on tissue oxygenation 
rather than just on ABG values. 

Critically ill patients who survive tend to have 
higher values for Do 2 and Vo 2 than do their counter- 
parts with similar illness who die. In an effort to in- 
crease survival rates, some investigators have ap- 
plied this observation by attempting to achieve the 
higher values of survivors in all patients. 49 " 51 De- 
spite enthusiastic reports from these groups, many 
authorities believe that both the rationale and the 
implementation of such a therapeutic approach de- 
serve further study before the approach can be 
widely applied in patient care. 3040 

One reason for the controversy surrounding 
pathologic supply-dependency of Vo 2 in critical ill- 
ness is the imperfect means available for assessing 
tissue oxygenation. Global measurements of Vo 2 
using so-called metabolic carts, although readily 
obtained in the ICU, are difficult to perform ac- 
curately and plagued by variations in the physio- 
logic baseline and other problems. 52 53 Calculation 
of Vo 2 from the Fick equation requires the presence 
of a pulmonary artery catheter and is subject to 
some of the same problems of interpretation in un- 
stable patients. Thus, there is a great need for non- 
invasive, readily available, accurate measures of 
global tissue oxygenation. 

Some clinicians believe that gastric tonometry 
may fill this need. 54 ' 55 With this technique, a saline- 
filled balloon is passed into the patient's stomach 
by means of a modified nasogastric tube. After a 
period of equilibration (typically 90 min), a sample 
of the saline is withdrawn and tested for Pco:- This 
Pco2 is believed to be essentially equivalent to gas- 
tric intramucosal Pco: and, thus, can be used to es- 
timate intramucosal pH using a simultaneously de- 
termined arterial-blood bicarbonate concentration 
and a modified version of the Henderson-Hassel- 
balch equation. 55 

Several studies 56 " 58 suggest that gastric intra- 
mucosal pH correlates with Do 2 , Vo 2 , the 2 - 
extraction ratio, mixed venous pH, and even the 
probability of survival and that this measurement 
therefore may be valuable in the assessment and 
management of critically ill patients. Further study 
will be required to determine the extent to which 




this index of gut mucosal hypoxia reflects whole- 
body oxygenation status, and hence will assist the 
clinician in managing acute respiratory failure. 59 

Acute Respiratory Failure as a Problem 
of Local Tissue Oxygenation 

Interest in measures of regional oxygenation as 
reflections of a patient's overall status indicates 
what may be a new direction in our thinking about 
acute respiratory failure. What once was con- 
sidered a disorder of pulmonary function and re- 
flected purely in ABG values is now thought of 
more globally as a problem of tissue oxygen supply 
and utilization. Clinicians are accustomed to envi- 
sioning this "tissue" as that of the body at large 
rather than of individual organs or perfusion beds, 
in part because only global measurements have 
been readily available. Yet Webster's tissue-based 
definition of respiration, 2 ' cited earlier, applies as 
well to individual tissue beds as to the whole or- 

The middle-aged airline passenger mentioned in 
Case 4 was concluded not to be in acute respiratory 
failure despite severe hypoxemia because his over- 
all Do 2 was normal. However, what if he had a pre- 
viously undetected narrowing in one coronary ar- 
tery, such that the hypoxemia produced by the 
airliner's abrupt loss of cabin pressure resulted in 
critical hypoxia of the myocardium served by that 
vessel? Although usually thought of as myocardial 
ischemia rather than hypoxia, at the tissue level it is 
the lack of 2 that produces the threat of cell injury 
and death. In this sense, cardiac ischemia, mes- 
enteric ischemia, or any process leading to acute 
tissue hypoxia, could be considered acute res- 
piratory failure; the step-down of P02 from alveolus 

Table 5. Historical Evolution of Concepts of Acute Respira- 
tory Failure 




1960s- 1970s Lungs 
1980s Whole body 

Arterial blood gas analysis 

Oxygen delivery/oxygen 

1990s Regional Gastric tonometry 

tissue beds ?MRI* ?others 

*MRI = magnetic resonance imaging. 

to respiring tissue depicted in Figure 7 varies with 
different tissue beds and metabolic needs. 

As mentioned earlier, clinicians assess and man- 
age patients according to prevailing concepts and 
use existing tools (Table 5). As technology ad- 
vances and means for assessing the oxygenation 
status of individual organs and tissue beds become 
available, the definition of acute respiratory failure 
may need to be revised once again, from an impair- 
ment in whole-body oxygenation to a description 
that includes local or regional problems. Such a 
definition might identify acute respiratory failure as 
acute life- or vital organ-threatening tissue hypoxia. 


1. Campbell EJM. Respiratory failure. Br Med J 1965:1: 

2. Bendixen HH, Egbert LD, Hedley-Whyte J, Laver MB, 
Pontoppidan H. Respiratory care. St Louis: CV Mosby. 

3. Petty TL. Intensive and rehabilitative respiratory care. 
Philadelphia: Lea & Febiger, 1971:3. 

4. Murray JF. Mechanisms of acute respiratory failure 
(Conference report: Division of Lung Diseases, Na- 
tional Heart, Lung, and Blood Institute). Am Rev Respir 
Dis 1977:115:1071-1078. 

5. Pontoppidan H, Geffin B, Lowenstein E. Acute res- 
piratory failure in the adult. N Engl J Med 1972:287: 

6. Pierson DJ. Respiratory failure: introduction and over- 
view. In: Pierson DJ. Kacmarek RM, eds. Foundations 
of respiratory care. New York: Churchill Livingstone, 

7. Pierson DJ. Acute respiratory failure. In: Sahn SA, ed. 
Pulmonary emergencies. New York: Churchill Living- 
stone, 1982:75-126. 

8. Marini JJ. Respiratory medicine and intensive care for 
the house officer. Baltimore: Williams & Wilkins. 1981: 

9. Pierson DJ. Clinical approach to the patient with acute 
ventilatory failure. In: Pierson DJ. Kacmarek RM. eds. 
Foundations of respiratory care. New York: Churchill 
Livingstone, 1992:707-719. 

10. Schmidt GA. Hall JB. Acute or chronic respiratory fail- 
ure: assessment and management of patients with 
COPD in the emergency setting. JAMA 1989:261:3444- 

1 1. Pepe PA, Hudson LD. Acute respiratory failure. In: Cal- 
laham ML. ed. Current practice of emergency medicine. 
Philadelphia: BC Decker, 1991:47-55. 

12. Derenne JP, Fleury B, Pariente R. Acute respiratory fail- 
ure of chronic obstructive pulmonary disease. Am Rev 
Respir Dis 1988:138:1006-1033. 




Hudson LD, Pierson DJ. Ventilatory failure. In: Pierson 
DJ. Kacmarek RM, eds. Foundations of respiratory care. 
New York: Churchill Livingstone, 1992:303-310. 
Aubier M, Murciano D. Milic-Emili J, Touaty E. Dagh- 
fous J, Pariente R. et al. Effects of the administration of 
O2 on ventilation and blood gases in patients with 
chronic obstructive pulmonary disease during acute res- 
piratory failure. Am Rev Respir Dis 1980;122:747-754. 
Hudson LD, Pierson DJ. Hypoxemia. In: Pierson DJ, 
Kacmarek RM, eds. Foundations of respiratory care. 
New York: Churchill Livingstone, 1992:311-318. 
Steinberg KP, Pierson DJ. Clinical approach to the pa- 
tient with acute oxygenation failure. In: Pierson DJ, 
Kacmarek RM, eds. Foundations of respiratory care. 
New York: Churchill Livingstone, 1992:721-739. 
West JB. State of the art: ventilation-perfusion rela- 
tionships. Am Rev Respir Dis 1977;1 16:919-943. 
West JB. Ventilation, blood flow, and gas exchange. In: 
Murray JF. Nadel JA, eds. Textbook of respiratory med- 
icine. Philadelphia: WB Saunders, 1988:47-84. 
Craig KC. Pierson DJ, Carrico CJ. The clinical applica- 
tion of positive end-expiratory pressure (PEEP) in the 
adult respiratory distress syndrome (ARDS). Respir 
Care 1985;30:184-201. 

Hall JB. Wood LDH. Oxygen therapy in the critically ill 
patient. In: Hall JB, Schmidt GA. Wood LDH, eds. Prin- 
ciples of critical care. New York: McGraw-Hill, 1992: 

Dantzker DR, Brook CJ, Dehart P, Lynch JP, Weg JG. 
Ventilation-perfusion distributions in the adult res- 
piratory distress syndrome. Am Rev Respir Dis 1979; 

Ralph D, Robertson HT. Respiratory gas exchange in 
adult respiratory distress syndrome. Semin Respir Med 

Webster's new international dictionary. Springfield 
MA: Merriam- Webster, 1981:1934. 
Taylor CR. Weibel ER. Design of the mammalian res- 
piratory system. 1. Problem and strategy. Respir Physiol 

Weibel ER. The pathway for oxygen. Cambridge MA: 
Harvard University Press, 1984. 

West JB. Ventilation/blood flow and gas exchange, 3rd 
ed. Oxford: Blackwell. 1977. 

Hudson LD, Pierson DJ. Inadequate blood oxygen con- 
tent. In: Pierson DJ. Kacmarek RM. eds. Foundations of 
respiratory care. New York: Churchill Livingstone, 

Hudson LD, Pierson DJ. Inadequate oxygen transport. 
In: Pierson DJ, Kacmarek RM. eds. Foundations of res- 
piratory care. New York: Churchill Livingstone. 

Pierson DJ. Preservation and augmentation of cardiac 
output. In: Pierson DJ, Kacmarek RM, eds. Foundations 
of respiratory care. New York: Churchill Livingstone. 

30. Hudson LD. Pierson DJ. Failure of peripheral oxygen 
extraction. In: Pierson DJ. Kacmarek RM. eds. Founda- 
tions of respiratory care. New York: Churchill Living- 
stone, 1992:327-330. 

31. Dolan MC. Carbon monoxide poisoning. Can Med As- 
soc J 1985;133:392-399. 

32. Jederlimc PJ. Irwin RS. Acute inhalation injury. In: 
Rippe JM. Irwin RS. Alpert JS. Fink MP. eds. Intensive 
care medicine, 2nd ed. Boston: Little. Brown & Co, 

33. Holland MA. Kozlowski LM. Clinical features and 
management of cyanide poisoning. Clin Pharm 

34. Hall AH, Rumack BH. Clinical toxicology of cyanide. 
Ann Emerg Med 1986:15:1067-1074. 

35. Hall AH. Systemic asphyxiants: cyanide and cyanogens, 
hydrogen sulfide, methemoglobin inducers. In: Rippe 
JM. Irwin RS. Alpert JS. Fink MP. eds. Intensive care 
medicine, 2nd ed. Boston: Little, Brown & Co, 1991: 

36. Cottrell JE, Casthely P. Brodie JD, Patel K. Klein A. 
Turndorf H. Prevention of nitroprusside-induced cya- 
nide toxicity with hydroxocobolamin. N Engl J Med 

37. Curry SC, Amold-Capell P. Toxic effects of drugs used 
in the ICU: nitroprusside, nitroglycerin, and angio- 
tensin-converting enzyme inhibitors. Crit Care Clin 

38. Samsel RW. Schumacker PT. Pathologic supply de- 
pendence of oxygen utilization. In: Hall JB. Schmidt 
GA, Wood LDH, eds. Principles of critical care. New 
York: McGraw-Hill, 1992:667-678. 

39. Shoemaker WC. Tissue perfusion and oxygenation: a 
primary problem in acute circulatory failure and shock 
states (editorial). Crit Care Med 1991;19:595-596. 

40. Russell JA. Oxygen delivery and consumption in adult 
respiratory distress syndrome and sepsis. Update In- 
tensive Care Emerg Med 1991;14:175-181. 

41. Weg JG. Oxygen transport in adult respiratory distress 
syndrome and other acute circulatory problems: re- 
lationship of oxygen delivery and oxygen consumption. 
Crit Care Med 1991;19:650-657. 

42. Cain SM. Curtis SE. Experimental models of pathologic 
oxygen supply dependency. Crit Care Med 1991:19: 

43. Edwards JD. Oxygen transport in cardiogenic and septic 
shock. Crit Care Med 1991:19:658-663. 

44. Danek SJ. Lynch JP. Weg JG. Dantzker DR. The de- 
pendence of oxygen uptake on oxygen delivery in the 
adult respiratory distress syndrome. Am Rev Respir Dis 

45. Tuchschmidt J. Oblitas D, Fried JC. Oxygen consump- 
tion in sepsis and septic shock. Crit Care Med 1991:19: 

46. Archie JP Jr. Mathematic coupling of data: a common 
source of error. Ann Surg 1 98 1 : 1 93:296-303. 

47. Samsel RW. Schumacker PT. Determination of the crit- 
ical O; delivery from experimental data: sensitivity to 
error. J Appl Physiol 1988:64:2074-2082. 





Ronco JJ, Phang PT. Walley KR, Wiggs B, Fenwick JC, 54. 

Russell JA. Oxygen consumption is independent of 

changes in oxygen delivery in severe adult respiratory 

distress syndrome. Am Rev Respir Dis 1991:143:1267- 55. 


Shoemaker WC. Appel PL, Kram HB, Waxman K, Lee 56. 

TS. Prospective trial of supranormal values of survivors 

as therapeutic goals in high-risk surgical patients. Chest 


50. Edwards JD, Brown GC, Nightingale P. Slater RM, Fa- 57. 
ragher EB. Use of survivors' cardiorespiratory values as 
therapeutic goals in septic shock. Crit Care Med 

51. Tuchschmidt J, Fried J, Astiz M. Rackow E. Elevation 58. 
of cardiac output and oxygen delivery improves out- 
come in septic shock. Chest 1992;102:216-220. 

Branson RD. The measurement of energy expenditure: 59. 

instrumentation, practical considerations, and clinical 
application. Respir Care 1990;35:640-659. 

Hess DR. Mundroff J. Assessment and monitoring of 60. 

metabolic and nutritional status. In: Pierson DJ, Kac- 
niarek RM. eds. Foundations of respiratory care. New 
York: Churchill Livingstone, 1992:541-554. 



Fiddian-Green RG. Should measurements of tissue pH 
and Po: he included in the routine monitoring of ICU 
patients? (editorial). Crit Care Med 1991:19:141-143. 
Fiddian-Green RG. Tonometry: theory and applications. 
Intensive Care World 1992;9:60-65. 
Gutierrez G, Palizas F. Doglio G. Wainsztein N. Gal- 
lesio A. Pacin J. et al. Gastric intramucosal pH as a ther- 
apeutic index of tissue oxygenation in critically ill pa- 
tients. Lancet 1992;339:195-199. 

Gutierrez G. Bismar H, Dantzker DR. Silva N. Com- 
parison of gastric intramucosal pH with measures of ox- 
ygen transport and consumption in critically ill patients. 
Crit Care Med 1992:20:451-457. 

Doglio GR, Pusajo JF. Bonfigli GC, Egurrola MA. Pa- 
lizas F. Gutierrez G. Predictive value of stomach mu- 
cosal pH in ICU patients (abstract). Chest 1990;98:54S. 
Silverman HJ. Gastric tonometry: an index of splanch- 
nic tissue oxygenation? (editorial). Crit Care Med 

Kacmarek RM. Oxygen carriage, transport, and utiliza- 
tion. In: Pierson DJ, Kacmarek RM, eds. Foundations of 
respiratory care. New York: Churchill Livingstone, 

Pierson Discussion 

Wood: David, that was really a nice 
historical review, but I ask you to 
consider two other things that are a 
part of this, to my way of thinking. 
The first is that the level of Vo : in the 
critically ill patient is often so much 
higher than normal (because of vari- 
ous aspects of hypermetabolism) that 
it plays a major role in what the Drji 
needs to be. Therapeutic adjustments 
of the V02 are a part of the game of 
treatment of oxygenation failure — 
therapeutic adjustments like de- 
creasing the work of breathing, or 
bringing down the temperature, or 
treating the sepsis. A related issue is 
that the Vo : and Do: influence arteri- 
al hypoxemia. You listed the five 
causes and put two of them in pa- 
rentheses. I think that we ought to 
add another one to those two in pa- 
rentheses, which is mixed venous hy- 
poxemia. On your slides, all the 
mixed venous blood has a P02 of 40 
torr and a content of 15 mL/dL. So 
often in critical illness, the PvO is 27 
and the content is way down, and 
that aggravates the amount of arterial 

desaturation for any degree of shunt 
or V/Q mismatch. 

Pierson: I agree with both of those 
comments and anticipate Jamie Stol- 
ler's paper on some of the manipula- 
tions you referred to about changing 
Vo 2 - It is true that I referred always to 
a normal baseline metabolic situation 
with these acute insults, but, as we all 
know, that's not the way it usually 
happens in the ICU. 

Stoller: While we're adding to the 
list of causes of hypoxemia, for the 
sake of completeness (but a little bit 
far afield from the ICU setting), we 
should mention the admixture of V/Q 
mismatching and anatomic shunt, so- 
called perfusion/diffusion impair- 
ment, that has increasingly come into 
the literature' 2 in the context of the 
hepatopulmonary syndrome. This 
syndrome, seen in patients with one 
of a variety of either acute or chronic 
liver diseases, seems to be associated 
with distention of the capillary bed 
and with a spongy characteristic on 
the pulmonary angiogram. The he- 
patopulmonary syndrome has been 
shown to be completely reversed by 

liver transplantation, including a re- 
port that we presented in Hepatology 
about a year ago.' Although it isn't 
always completely reversed, it can 
be. Again, though this is a little bit 
far afield from the ICU setting on 
which we're focusing, I think people 
are increasingly adding this to the 
list of the five or six causes of hy- 

1. Krowka MJ. Cortese DA. Pulmonary 
aspects of chronic liver disease and 
liver transplantation. Mayo Clin Proc 

2. Stanley NN, Woodgate DJ. Mottled 
chest radiograph and gas transfer de- 
fect in chronic liver disease. Thorax 

3. Stoller JK. Moodie D. Schiavone 
WA. Vogt D. Broughan T, Winkel- 
man E, et al. Reduction of intra- 
pulmonary shunt and resolution of 
digital clubbing associated with pri- 
mary biliary cirrhosis after liver 
transplantation. Hepatology 1990;11: 

Pierson: Those comments are not so 
far afield from our discussion of hy- 
poxic respiratory failure. I have cer- 
tainly seen patients with severe liver 




cirrhosis, portal hypertension, chron- 
ic gas exchange derangements, and a 
kind of physiologic shunting, who 
become critically ill. The clinicians 
at the bedside are waiting for their 
arterial oxygenation to improve 
enough to start weaning, but such pa- 
tients will never get there because 
their baseline oxygenation is so em- 

Nelson: As Larry (Wood) men- 
tioned, I think it's very important 
that add to in your list of five causes 
of hypoxemia, reduced cardiac out- 
put and all causes of venous oxygen 
desaturation. I think that in our in- 
tensive care unit (as in many places 
that deal with very acutely ill pa- 
tients) probably venous desaturation, 
with some element of low V/Q or 
shunt, is another cause of hypox- 
emia. This has caused me, when I 
teach pathophysiology and hypox- 
emia, to change John West's classic 
list of six things and divide it. 1 Now I 
have oxygen-responsive hypoxemia, 
the four classic things that you in- 
dicated (low V/Q. low Pio : , hypo- 
ventilation, and diffusion) with diffu- 
sion and Pio: in parentheses on my 
slide, and then the oxygen-refractory 
ones (Qsp/Qt and low Sy0 2 )- This 
gives you diagnostic information, it 
starts your therapy, and it lets you 
know that those oxygen-refractory 
causes of hypoxemia — shunt and low 
Syo 2 states — are the bad ones. The 
latter are the high-mortality, high- 
intensity of care states and the for- 
mer low-mortality, low-intensity of 

1. West JB. Gas exchange. In: Pul- 
monary physiology — the essentials, 
3nd ed. Baltimore: Williams & Wil- 
kins, 1987:19-41. 

Pierson: Those mechanisms still rely 
on the same basic processes that I 
showed. If you and I had run the 
100-yard dash and our mixed venous 
P02 (PvOi) had gone way, way down, 
we wouldn't be hypoxemic at the 

end because we don't have lung re- 
gions with very low V/Q or shunt. 
We. as you mentioned, have to have 
the predisposing presence of one of 
those fundamental derangements. 

Hudson: I want to emphasize one 
cause of oxygenation failure that you 
alluded to that I think we can ma- 
nipulate — the affinity of the oxygen 
for hemoglobin (Hb). The example 
that you gave of carbon monoxide 
poisoning, I think, is a good example 
because all of us know that patients 
who have a severe anemia with a Hb 
level that is half their normal value 
can do quite well, but patients with 
carbon monoxide poisoning with a 
comparable level of available Hb 
(that is, half their Hb unbound by CO 
and thus presumably available for 
carrying oxygen) have profound 
problems clinically. That's probably 
due to the shift of the curve to the 
left, so that the oxygen is more tight- 
ly bound. One of the things that con- 
trols that is pH. and I think we some- 
times overlook that when we're 
looking at blood gases and ven- 
tilating patients so that we are low- 
ering their CO : . We may have them 
in an alkalotic range when one of the 
simple things we could do is just let 
them be mildly acidotic and probably 
improve that oxygen delivery. The 
other thing I'd like to ask about — we 
always put diffusion limitation in pa- 
rentheses, indicating that it is not im- 
portant clinically (as you have, and I 
have as well), but I'm wondering 
about it in surgical and trauma pa- 
tients who have cardiac outputs four 
or five times normal, or three times 
normal, anyway. Might diffusion not 
be a limiting factor or at least a con- 
tributing one in this situation? I think 
it may be less important because I'm 
not sure what we would do about it. 
But I'm wondering about it. 

Pierson: At the 1982 conference, 
John Murray gave this talk, 1 and I 
asked him that question (as I dis- 
covered last night in reviewing the 

discussions after those articles). He 
said that the same physiology still 

1 . Murray JF. Pathophysiology of acute 
respiratory failure. Respir Care 1983; 


Hudson: I thought that Wagner and 
his colleagues, when they had looked 
at various states, 1 " 1 found that if you 
were exercising, it could become a 
limiting factor — that seems to me to 
be at about the same level of cardiac 
output. Larry (Wood) or Frank (Cer- 
ra). do you have any comments? 

1. Gale GE, Torre-Bueno JR. Moon RE. 
Saltzman HA, Wagner PD. Ventila- 
tion-perfusion inequality in normal 
humans during exercise at sea level 
and simulated altitude. J Appl Physiol 

2. Torre-Bueno JR, Wagner PD, Saltz- 
man HA. Gale GE. Moon RE. Diffu- 
sion limitation in normal humans dur- 
ing exercise at sea level and simulated 
altitude. J Appl Physiol 1985:58:989- 

3. Hammond MD. Gale GE. Kapitan 
KS, Ries A, Wagner PD. Pulmonary 
gas exchange in humans during ex- 
ercise at sea level. J Appl Physiol 

Cerra: I have three comments to get 
at some of the questions you were 
raising. The first is. I think we have 
to accept that in tissue-injury states 
or inflammatory states, oxygen de- 
mand is increased. The problem we 
have is that we have no way of di- 
rectly measuring what that demand 
is. We infer it by looking at supply- 
dependency. There are two other di- 
mensions. One is the hypothesis re- 
lated to flow distribution and per- 
fusion versus diffusion (as has been 
mentioned). I think that your analysis 
of where we're going with these def- 
initions is right on target and that, ul- 
timately, we're going to end up in a 
changing position. Whereas we've 
been 'sitting' at the heart-lung com- 
plex looking out. we're now going to 




'sit' on the cells and look at what is 
happening and being delivered there. 
The second dimension is that we 
have to consider cell energetics and 
metabolism. They are really unre- 
lated to oxygenation — more a func- 
tion of the mediators of the in- 
flammatory response. You have to 
get into that controversy because un- 
less we're actually monitoring cell 
energetics and metabolism, it's dif- 
ficult to figure out where some of the 
studies fit in. In other words, are we 
dealing with an anaerobic problem or 
a different form of aerobic energy 
production? These considerations 
can really change the meaning of a 
lot of the indicators that we're cur- 
rently using. 

Wood: I want to reply to Len's 
(Hudson) question and then ask him 
a question or make a comment on 
something that he raised. Sitting on 
the cell looking back, as Frank Cerra 
said, gives a value of Po: of about 1 
torr that is associated with anaerobic 
metabolism. If you look back to the 
capillary end, even in the high-flow 
states that Len (Hudson) asked 
about, you get to a value of about 5-6 
ton that causes anaerobic me- 
tabolism. In the mixed venous blood 
of all of our anaerobic patients, PyO : 
isn't very much below 27 torr when 
the patient begins to get lactic acidot- 
ic. Where is the difference? Some of 
the excitement in looking at this is 
that in order for the cells to have 
gone hypoxic at a level of 5-6 torr, 
yet PvO: 1S 27, there must be a big 
peripheral shunt out there some- 
where that accounts for the mixed 
venous blood coming back so high. 
So, I think that the question needs to 
be answered concerning hypermet- 
abolic patients — surgical, even pneu- 
monia patients — with 10, 12, 15L/min 
cardiac output, who have lactic ac- 
idosis: Is this anaerobic metabolism 
due to insufficient Do:, or what? 
There's a lot to be understood about 
when the oxygen failure that David 

Pierson talked about is in fact the 
cause of anaerobic metabolism. An- 
other area of excitement for me is the 
issue that you raised — let the patient 
be a little more acidotic because eve- 
rybody knows that carbon monoxide 
in your patient who's dragged from 
the burning building left-shifts the 
oxyhemoglobin dissociation curve, 
and that's bad. But, all the high al- 
titude animals have left-shifted oxy- 
hemoglobin dissociation curves. They 
do that because when there are hy- 
poxic alveoli, arterial saturation is in- 
creased by a left shift. Such a left- 
shift is also very good for our very 
hypoxic patients because it gets their 
hemoglobin more loaded with Oi, 
provided they can offload O; in the 
tissues. There are now two studies in 
the literature in the last 5 years, one 
by Paul Schumacker from our group, 
who left-shifted the oxyhemoglobin 
dissociation curve in experimental 
animals and then followed the Vor 
Do: relationship. 1 2 He found that the 
critical extraction fraction didn't 
change, but the PyO : at which the an- 
imals became anaerobic fell from 25 
to 15. In other words, the tissues ex- 
tract the amount of oxygen that they 
need to maintain metabolism down 
through 25, 20, right down to 15; so, 
PvCh is almost irrelevant. What he 
also found was that there was no im- 
pediment to offloading, with a left- 
shift of the oxyhemoglobin dis- 
sociation curve. What we do, when 
push-comes-to-shove in a very hy- 
poxic patient, is lower the F102 to 
limit lung O: toxicity, hyperventilate 
the Pco; to 25 torr to left-shift the oxy- 
hemoglobin dissociation curve, and 
increase arterial saturation and Do;, 
and hope that these animal studies 
are relevant to the critically ill pa- 
tient's ability to extract O: in the tis- 

1. Schumacker PT, Rowland J, Saltz S, 
Nelson DP, Wood LDH. Effects of 
hyperthermia and hypothermia on ox- 

ygen extraction by tissues during hy- 
povolemia. J Appl Physiol 1987:63: 
2. Schumacker PT, Long GR. Wood 
LDH. Tissue oxygen extraction dur- 
ing hypovolemia: role of hemoglobin 
P50. J Appl Physiol 1987:62:1801- 

Pierson: I'd like to comment on 
your remark about the very low P02 
in the tissues at which we get into 
trouble, and how different that num- 
ber is from the PvO values we re- 
gard as indicating a critical threat of 
organ damage or death in our ICU 
patients. I was thinking about that 
difference as we were flying down 
here. As we crossed over Baja Cal- 
ifornia, there was a place where a 
number of dry tributaries joined into 
one stream. It seems to me that for a 
long time we've been sampling the 
river downstream, after all the in- 
dividual tissue streams have come to- 
gether, and we're saying that that's 
what is happening in the patient as a 
whole. That's what we're doing with 
the PvO- Now. with gastric tonom- 
etry and other new techniques, we're 
going up one of those streams and 
hoping that by sampling up there 
we'll get a more accurate image of 
what's going on in the tissues. Clear- 
ly, the critical tissue hypoxia you re- 
fer to, in which the P02 is 5 or down 
to 1 torr, may be a relatively local- 
ized phenomenon somewhere that 
will be diluted a great deal by 
streams from a lot of other tributaries 
that aren't so severely compromised. 
I agree with you that we need the 
ability to measure what's going on in 
the relevant tributary — neither in a 
random tributary nor in the river way 
downstream after they've all mixed. 

Wood: Just to put it out there for 
everybody to think about — you need 
a 50% intraorgan shunt to explain 
those numbers. In other words, 50% 
of the blood flow going through the 




anaerobic organ has to go past the tis- 
sues, giving up no oxygen, in order 
to raise the PyCb coming out of that 
organ from the anaerobic tissues (5-6 
torr) to a partial pressure of 27 torr. 
That's not what people have found 
when they go looking for peripheral 
anatomic shunts, so shunt gets to be a 
tricky word here. There can be V02- 
Do; variance and counter-current dif- 
fusion that can account for the high 
PvO- too; but one way or the other — 
counter-current diffusion or Vo:-Do2 
variance, analogous to the V/Q var- 
iance in the lung — the effective pe- 
ripheral shunt has to be on the order 
of 50% to account for anaerobic me- 
tabolism and an organ venous Po; of 

Pierson: But that's assuming that the 
whole tissue out there is behaving the 
same and that what you're sampling 
in the mixed venous blood reflects 
homogeneity out in the periphery, 
doesn't it? 

Wood: The Voj-Do: mismatch must 
be among organs or within organs or 

Cerra: I would like to amplify one 
of the comments that Larry Wood 
made. We need to be looking at not 
only the energetics of cells but also 
the total metabolic responses of those 
cells. We can't evaluate this by look- 
ing at purely physiologic parameters. 
What you're calling a physiologic 
shunt in those microcirculatory paths 
does not meet the metabolic defini- 
tion of shunt. We've got to integrate 
those two concepts in order to re- 
solve this issue. I will be reviewing 
this metabolic data in my presenta- 

Hudson: I have one last comment to 
Larry (Wood). I've learned some- 
thing from what you've just said. But 
one of the things that I think it repre- 
sents is that I was speculating and 
dealing from theory, whereas you ac- 
tually studied the situation and pro- 

vided data. I think that's an im- 
portant thing we should be trying to 
do here. I wonder if the same situa- 
tion you studied in the animal mod- 
els is not approachable in patients, 
because this seems to be an im- 
portant issue, and one that could be 
modified in practice. Do you think 
that is approachable, in terms of be- 
ing able to study that point in the 
clinical situation? 

Wood: Yes, it is approachable. It 
can be studied. It's a tough study in 
animals, if you're referring to left- 
shifting the oxyhemoglobin dissoci- 
ation curve. But a lot of tough stud- 
ies are done, and as we get better at 
defining the relevant mechanisms in 
animal studies, we get better at the 
clinical investigation of critically ill 
patients; so, yes, it can, and I think it 
should be done. Push-comes-to-shove 
regarding limits of aerobic metabo- 
lism more often these days than it 
used to. so there are good clinical in- 
vestigations to be done. 

AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibhio 


American Association 
for Respiratory Care 
39th Annual Convention 
and Exhibition 

Nashville, Tennessee 



AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition ■ AARC 39th Annual Convention and Exhibition 



Clinical Effects of Hypoxemia and Tissue Hypoxia 

Thomas L Higgins MD and Jean-Pierre Yared MD 

I. Introduction 
II. Definitions 

III. Physical Effects, Symptoms, & Signs of Hypoxemia 

A. Central Nervous System Effects 

B. Effects Seen in Sleep Apnea 

C. Chemoreceptor-Mediated Response 

D. Cardiovascular Effects 

E. Pulmonary Effects 

F. Effects on Splanchnic & Renal Circulation 

G. Gastrointestinal Effects 
H. Renal Effects 

I. Fetomaternal Effects 

IV. Hypoxia Secondary to Decreased Oxygen Delivery 

V. Hypoxia Secondary to Decrease in Oxidative Metabolism 
VI. Limits of Tolerance to Hypoxemia 
VII. Hypoxemia in Clinical Practice 

A. Desaturation in Special Clinical Settings 

B. Reliability of Signs & Symptoms 

C. Reliability of Oximeter Alarms 

D. Differential Diagnosis of Hypoxemia & Its Management 
VIII. Summary & Conclusions 


Oxygen is the energy source required by all organ 
systems for proper function. Hypoxia is the final 
common pathway for most, if not all, cardio- 
respiratory events resulting in morbidity and mor- 
tality. This chain of events includes inadequate sup- 
ply of inspired oxygen, deficiencies of ventilation 
and perfusion, and insufficient oxygen delivery to 

Dr Higgins and Dr Yared are associated with the Department 
of Cardiothoracic Anesthesia and Cardiothoracic Intensive 
Care, The Cleveland Clinic Foundation. Cleveland. Ohio. 

A version of this paper was prepared for presentation at the 
Respiratory Care Journal Conference on Oxygenation in the 
Critically 111 Patient, held in Puerto Vallarta, Mexico, October 
8-10, 1992. 

Reprints: Thomas L Higgins MD, Director, Cardiothoracic In- 
tensive Care Unit, The Cleveland Clinic Foundation. 9500 Eu- 
clid Ave G5-164. Cleveland OH 44195-5080. 

and utilization by the tissues. The importance of 
monitoring for decreased arterial oxygen content 
(hypoxemia) and tissue hypoxia has long been rec- 
ognized. However, until the advent of oximetry, di- 
rect detection of hypoxemia was delayed because 
of the intermittent nature of blood gas measure- 
ments, whereas indirect detection of arterial de- 
saturation was unreliable because of the lack of 
sensitivity and specificity of the clinical signs as- 
sociated with hypoxemia. As a result, monitoring 
adequacy of tissue and organ oxygenation is often 
delayed and difficult to achieve. However, the 
brain and the heart are two organs in which the di- 
agnosis of hypoxia can be made soon after the on- 
set of inadequate oxygen delivery: Cerebral hypox- 
ia in the conscious patient can be diagnosed early 
by clinical observation of the neurologic status, 
whereas myocardial ischemia can be detected by 
assessing the electrocardiogram, the hemodynamic 
profile, and the echocardiogram. Unfortunately, it 




1 . Low Inspired Oxygen 

2. Impaired Ventilation 




4. Venous-Arterial Shunts 

. 5 Impaired Circulation 

. _ 6 Decreased Carrying Capacity 
7 Increased Consumption 

r^ r\ f^ 8 Abnormal Utilization 

Fig. 1 . The causes of hypoxia. 

is often necessary to wait for tissue damage to oc- 
cur before evidence of end-organ dysfunction can 
be seen when oxygen delivery to the kidney, liver, 
or gastrointestinal tract is impaired. Because the 
goal of therapy is prevention of tissue hypoxia rath- 
er than mere correction of a low arterial oxygen 
tension, the response to detection of hypoxemia 
varies in magnitude and urgency, depending on the 
associated symptoms, physical signs, and phys- 
iologic effects. This article focuses on hypoxemia, 
its clinical effects on various organ systems, and its 
incidence during sleep and during common clinical 
situations. Monitoring techniques, differential di- 
agnosis, and therapy of hypoxemia and hypoxia are 


Although both hypoxemia and hypoxia are de- 
fined as a subnormal oxygen content, the former re- 
fers to the arterial blood while the latter refers to or- 
gans and tissues.' The terms anoxia and hypoxia 
have become synonymous, despite the common 
clinical usage that implies a more complete oxygen 
lack when the term anoxia is used. The differential 
diagnosis of hypoxia 232 * :s " encompasses hypotonic 
anoxemia (diminution of arterial oxygen tension in a 
normal individual due to low inspired oxygen or 
high altitude), anemic hypoxia or isotonic anoxemia 
(reduced effective red cell mass or abnormal hemo- 

globin), stagnant hypoxia or hypokinetic anoxia 
(generalized or regional decrease in blood supply as 
in low cardiac output states), and histotoxic anoxia 
(poisoning of the oxidative enzymes) as with cya- 
nide poisoning. This latter condition is not truly hy- 
poxemia because the content of the arterial blood is 
normal, but it mimics hypoxemia by preventing cel- 
lular utilization of oxygen via disruption of mito- 
chondrial oxidative metabolism (Fig. 1 & Table 1 ). 

Physical Effects, Symptoms, and 
Signs of Hypoxemia 

Hypoxemia rarely exists in isolation, due in part 
to accompanying hypercapnia and acidosis in many 
illnesses, and in part due to the normal com- 
pensatory responses. The setting of a high altitude 
environment is useful in isolating the effects of hy- 
poxemia alone because the partial pressure of oxy- 

Table 1. Causes of Hypoxia and Associated Conditions or 


Condition or Circumstance 

Low F102 

Impaired ventilation 

Altitude, closed space, iatrogenic error 
(error in setting Fio:) 

Airway obstruction, neuromuscular 
failure, depressed respirators dri\e. 
pneumothorax, flail chest, broncho- 

Impaired oxygenation Pulmonary edema, pneumonia, pulmo- 
nary hemorrhage, drowning, intersti- 
tial luns disease 

Venoarterial shunt 

Impaired circulation 

Decreased oxygen 
content of arterial 

Increased oxygen 

Abnormal oxygen 

V/Q mismatch, congenital heart dis- 
ease, acquired ventricular septal 

Congestive heart failure, shock, hem- 
orrhage, sepsis, thrombosis (em- 
bolus, abnormal distribution of 
blood flow) 

Anemia, carbon monoxide poisoning, 
abnormal hemoglobin (eg. methemo- 

Hypermetabolic states; fever, sepsis, 
post-surgery, malignant hyperther- 
mia, hyperthyroidism 

Inhibition of cellular metabolism 
(eg. cyanide poisoning) 




gen decreases directly with decreases in barometric 
pressure. The relative composition of atmospheric 
air or fraction of inspired oxygen (Fio:) does not 
change with altitude, but at 18,000 ft barometric 
pressure is about 380 torr, half that at sea level, 2 
and the inspired partial pressure of oxygen (P102) is 
reduced to about 70 torr (Pio 2 = 0.21 [380 - 47], 
where 47 torr is the vapor pressure of water at body 
temperature). The resultant alveolar partial pressure 
of oxygen (Pao:) in turn decreases to about 33 torr. 
A prolonged stay at high altitude mimics chronic 
hypoxemia and results in a compensatory increase 
in red blood cell mass, hemoglobin concentration, 
and blood volume, and in respiratory alkalosis. In 
contrast, acute hypoxemia can be simulated by rap- 
id ascent. Mountain climbing expeditions to study 
the effects of hypoxemia were first conducted in 
the 1890s. However, it was recognized that the 
time necessary for the ascent allowed some degree 
of physiologic readjustment to occur. Ascent in an 
unpressurized airplane without supplemental oxy- 
gen is a more useful model for study because 
changes occur more rapidly. These studies show 
that at 1 8,000 ft, an Fio : of about 0.45 is needed to 
achieve 95% saturation of arterial blood in normal 
subjects. At 34,000 ft, an F102 of 1.0 is required to 
achieve 95% saturation, and it follows that arterial 
hypoxemia develops in unpressurized cabins at 
higher altitudes, even when pure oxygen is sup- 
plied. A rapid airplane ascent made without pres- 
surization or supplemental oxygen causes the sub- 
ject to experience sensations of excitement, exhila- 
ration, and well-being. As altitude increases, more 
serious problems develop, including mental and 
sensory dullness, muscular weakness, headache, 
vomiting, dyspnea, cyanosis, periodic breathing, 
and a tendency to develop fixed ideation resulting 
in inappropriate or dangerous behavior. 2 

Conditions at varying altitudes can be simulated 
by decreasing ambient pressure in an airtight cham- 
ber, a model that can be modified to produce yet 
more rapid changes. Rapid onset of hypoxemia pro- 
duces sudden loss of consciousness. At a simulated 
altitude of about 20,000 ft, with the inspired oxy- 
gen tension (P102) 'ess than half the tension at sea 
level, most subjects experience failing vision, 
incoordination, and inability to perform simple 
mental tasks. If an altitude of 26,000 ft is simulat- 
ed, no subject is able to remain conscious for long- 
er than 15 minutes. 2 

Central Nervous System Effects 

The central nervous system is the organ system 
most sensitive to hypoxemia, probably because 
integration of signals involving multiple synapses 
is quick to demonstrate the cumulative effects of 
slight performance degradation at each level, 
whether caused by hypoxemia or drugs such as al- 
cohol. In clinical practice, early neurologic findings 
associated with hypoxemia include dimmed pe- 
ripheral vision and impaired mental and psycho- 
motor performance. Other common findings are 
sleeplessness, delirium, drowsiness, forgetfulness. 
irritability, restlessness, confusion, and inatten- 
tiveness. Seizures and unconsciousness are hall- 
marks of more profound hypoxemia. 2 During hy- 
poxemia, the body responds by increasing cerebral 
blood flow through cerebral vasodilatation and aug- 
mentation of cardiac output. 333 351 If the increase in 
cerebral blood flow is insufficient to promote a nor- 
mal oxygen delivery, the ATP-dependent sodium 
pumps may fail, allowing sodium and water to ac- 
cumulate within cells and resulting in cerebral ede- 
ma. 4 In patients with chronic hypoxemia, poly- 
cythemia is a common compensatory mechanism. 
Although the increase in red cell mass and hemo- 
globin increases oxygen-carrying capacity, it also 
increases blood viscosity, causing an important de- 
crease in cerebral blood flow that can offset the 
beneficial effect of the increase in arterial oxygen 
content (C a o:)- 5 

The electroencephalogram (EEG) demonstrates 
important changes with hypoxemia, although in the 
critically ill these changes may be difficult to separ- 
ate from the effects of drugs or concurrent meta- 
bolic derangements. Within 20 seconds from the 
onset of acute anoxemia, the EEG becomes iso- 
electric, but it is reversible if the oxygen supply is 
resumed within 8 minutes.* 337 Hypoxemia is in- 
itially associated with moderate slowing of the 
EEG rhythm, and then, as the oxygen deficit be- 
comes more severe, the appearance of delta waves. 
With extreme hypoxemia, there is a generalized de- 
cline in amplitude of the EEG, which eventually 
becomes flat. Generalized periodic patterns, such 
as burst suppression, carry a poor prognosis. 6 

Respiratory insufficiency has important effects 
on neurologic function. Those effects are often due 
to a combination of hypoxemia and hypercapnia 




because both disturbances usually coexist. Hyper- 
capnia has additive effects to hypoxemia, leading 
to more vasodilatation and an increase in cerebral 
blood flow, and ultimately resulting in further in- 
crease in intracranial pressure. 7 

Effects Seen in Sleep Apnea 

During the rapid eye movement ( REM ) phase of 
normal sleep, the breathing pattern is irregular. The 
response to hypercapnia is inhibited during REM 
sleep, but the response to hypoxemia is retained. 8 
However, in patients with sleep apnea syndrome, 
the response to hypoxemia is blunted. 9 Obstructive 
sleep apnea, characterized by repetitive episodes of 
upper airway obstruction during sleep, is associated 
with hypoxemia and sleep disruption. Other clin- 
ical correlates include daytime hypersomnolence, 
cognitive impairment, systemic and pulmonary hy- 
pertension, and cardiac arrhythmias. Patients with 
severe obstructive sleep apnea can develop hypox- 
emia (with oxygen saturations falling to < 75%). 10 

Sleep apnea induces major changes in the car- 
diovascular system. Repetitive, episodic hypox- 
emia is associated with chronic hypertension: Ex- 
perimental animals subjected for 30 consecutive 
days to periods of hypoxemia simulating sleep ap- 
nea (multiple 30-second periods at an F102 = 0.03- 
0.05 for 7 hours/day) developed both hypertension 
and increased left ventricular size." Simulation of a 
sleep-apnea breathing pattern in healthy volunteers 
caused periodic hypertension in healthy volunteers, 
with the end-apneic blood pressure correlating in- 
versely with arterial oxygen saturation. 12 

Chemoreceptor-Mediated Response 

Chemoreceptors are located in the carotid bodies 
at the bifurcation of the common carotid arteries 
and in the aortic bodies found in the arch of the 
aorta. They have a very large blood supply that re- 
sults in minimal oxygen extraction — thus exposing 
the chemoreceptor cells to blood with pH. P02. and 
PC02 similar to those in arterial blood. 

Carotid body chemoreceptors respond to hypox- 
emia by increasing the respiratory drive, thus in- 
creasing minute ventilation. They appear to be lo- 
cated in the afferent nerve terminals. 13 The ability 
of the peripheral dopamine-antagonist (D : ) dom- 
peridone to augment the hypoxic and hypercapnic 

ventilatory responses supports the concept of dopa- 
minergic modulation of the activity of the carotid 
body. 14 The glomus cells appear to maintain a tonic 
dopaminergic inhibition of the afferent impulses. 
Chemoreceptor activity increases with decrease in 
P a o;. However, when P a o: is maintained at less 
than 30 torr for a prolonged period, the increased 
activity cannot be sustained. These chemoreceptors 
initiate a powerful respiratory drive that continues 
to operate, even when the respiratory centers stop, 
and to respond to direct chemical activity. This is 
why they play an important role in sustaining res- 
piration in patients with chronic C0 2 retention in 
whom judicious oxygen therapy is important to 
prevent further respiratory depression. 

Cardiovascular Effects 

The initial effects of hypoxemia on the car- 
diovascular system are tachycardia, increased 
stroke volume, and hypertension (a reflex initiated 
via feedback from carotid body chemoreceptors). 
Oxygen extraction by the myocardium is nearly 
maximal, and any decrease in oxygen-carrying ca- 
pacity or increase in oxygen demands has to be met 
by an appropriate increase in coronary blood flow. 
As arterial saturation is experimentally decreased 
from 95% to 64%, coronary blood flow increases 
to maintain constant myocardial oxygen consump- 
tion. When coronary blood flow is maximal, fur- 
ther decreases in arterial saturation cannot be com- 
pensated for. and myocardial ischemia results. The 
electrocardiogram shows S-T segment elevation, 
and biochemical measurements show a decrease in 
lactate extraction, indicating a shift from aerobic to 
anaerobic metabolism. 15 Severe hypoxemia is asso- 
ciated with depression of myocardial contractility 
and development of dysrhythmias, followed by 
bradycardia, hypotension, and cardiac arrest. Hy- 
poxemia has also been implicated as a major factor 
determining the severity of bupivacaine toxicity: A 
PaO: < 40 torr enhances cardiotoxicity and neuro- 
toxicity and increases the likelihood that arrhyth- 
mias will occur before seizures. u> 

Pulmonary Effects 

Alveolar oxygen tension is a major determinant 
of pulmonary vascular tone, which in turn regulates 




distribution of pulmonary blood flow. Von Euler 
and Liljestrand, 17 in 1946, were the first to demon- 
strate experimentally that alveolar hypoxia in- 
creases pulmonary artery pressure, a response that 
is enhanced by coexisting hypercapnia or extra- 
cellular acidosis. Vasoconstriction in response to 
hypoxemia is unique to the pulmonary circulation, 
contrasting markedly with the vasodilation ob- 
served in the systemic circulation. Systemic vessels 
dilate in an attempt to provide more blood flow to 
the hypoxic organ, but the pulmonary vasculature 
constricts to diminish perfusion of underventilated 
alveoli and thus diverts pulmonary blood flow to 
better ventilated areas and maintains appropriate 
ventilation-to-perfusion matching. Normally, vaso- 
constriction occurs at a localized level, affects only 
the poorly perfused alveoli, and, thus, allows over- 
all pulmonary vascular resistance and pulmonary 
artery pressure to remain constant. With gener- 
alized hypoxia, as it occurs in individuals living at 
high altitude, the vasoconstrictive response is gen- 
eralized, resulting in chronic irreversible pul- 
monary hypertension. The process is characterized 
by hypertrophy of the muscularis of small pul- 
monary vessels and development of a muscle layer 
in other vessels that were originally nonmuscular. 18 
The increase in pulmonary vascular resistance can 
in turn lead to right ventricular failure. 

The mechanism of this unique vasoconstrictor 
response to hypoxia in the pulmonary vasculature 
is not yet fully elucidated. Pulmonary vasocon- 
striction results from either airway or vascular hy- 
poxia. Gases from the airway can diffuse directly 
into large pulmonary arterioles, 19 and it is possible 
that cells in the precapillary pulmonary arteriole 
sense oxygen tension and initiate vasoconstriction 
via a mechanism that is still obscure. 2021 Hypoxia, 
however, has a direct relaxing effect on isolated 
strips of pulmonary artery similar to the effect on 
other smooth muscles. 22 It is also possible that bio- 
chemical mediators (including noradrenaline, 23 his- 
tamine, 24 prostaglandins, 23 2<> nitric oxide, 27 and cal- 
cium 28 ) play a role in mediating this response. Such 
evidence comes from studies evaluating the pulmo- 
nary vascular response to mediators and their an- 
tagonists. In clinical practice, it is important to ap- 
preciate the potential for drugs to inhibit hypoxic 
pulmonary vasoconstriction. Such drugs include ni- 
troglycerin, 29 nitroprusside. 29 3 " nifedipine, 2831 iso- 
proterenol, dobutamine, 32 nitric oxide, 27 and several 

anesthetic agents. The inhibitory effect of inhaled 
anesthetics is addressed later. 

Effects of Hypoxia on Splanchnic and Renal 

Hypoxemia affects the function of intra-ab- 
dominal organs both directly and indirectly. Hy- 
poxemia is associated with an increase in sym- 
pathetic tone that causes vasoconstriction and re- 
duces blood flow to the kidneys, liver, gall bladder, 
stomach, gut, and pancreas. This effect is potentiat- 
ed by coexisting hypercapnia and by the neuro- 
endocrine response to stress. 33 Hypoxia can also 
result from a decrease in the perfusion of these or- 
gans secondary to low cardiac output states during 
which blood flow is diverted to more vital organs. 

Gastrointestinal Effects 

Hypoxic injury to intra-abdominal organs ap- 
pears to play a major role in the pathogenesis of the 
multiple organ systems failure that often accom- 
panies critical illness, affecting outcome adversely. 
Stress-related gastric mucosal injury resulting in se- 
vere bleeding carries a poor prognosis and is 
thought to result from mucosal ischemia. Ischemia 
impairs the function of the gastric epithelial cells, 
the production of prostaglandins, the integrity of 
the protective mucopolysaccharide layer that coats 
the gastric mucosa, and the transport of bicarbonate 
that is needed to buffer the hydrogen ions that pen- 
etrate the mucosal barrier. As a result, the mucosa 
becomes vulnerable and ulceration ensues. 3435 Sim- 
ilar mechanisms have been shown to affect the bow- 
el, liver, gall bladder, and pancreas. 36,37 Splanchnic 
ischemia has been implicated as the initiating force 
of multiple organ systems failure and sepsis. How- 
ever, not all patients who show evidence of splanch- 
nic ischemia develop secondary injury such as pan- 
creatitis, erosive gastritis, intestinal infarction, 
cholecystitis, or hepatic failure. Ischemia has been 
detected in about 50% of patients undergoing car- 
diac surgery, but only 1.1% of them develop signs 
of gastrointestinal disease; 38 however, when such a 
complication occurs, mortality is high. 3(1 

Renal Effects 

When oxygen supply to the kidney decreases to 
less than two thirds of normal, the kidney's ability 




to handle excretion of urea and sodium is impaired. 
With more severe hypoxia, tubular necrosis oc- 
curs. 321 "' In addition to the decrease in renal 
blood flow associated with hypoxemia, a decrease 
in oxygen tension also affects an important meta- 
bolic function of the kidney. The kidney is the site 
of synthesis of a renal erythropoietic factor, which 
cleaves part of a serum substrate to produce func- 
tional erythropoietin. 39 Both acute and chronic hy- 
poxemia stimulate the production of the hormone, 
inducing an increase in production of red blood 
cells and hemoglobin. 

Fetomaternal Effects 

In utero, fetal oxygen supply depends on an ad- 
equate uterine blood flow and oxygen content, both 
of which can be compromised during pregnancy, 
particularly at term and during labor. The pressure 
exerted by the gravid uterus on the inferior vena 
cava in the supine position can significantly de- 
crease venous return, cardiac output, blood pres- 
sure, and uterine blood flow. In addition, changes 
in maternal pulmonary physiology, particularly dur- 
ing the third trimester of pregnancy and labor, are 
associated with a higher incidence of maternal and 
fetal hypoxemia: The mother's functional residual 
capacity decreases because of the cephalad dis- 
placement of the diaphragm by the gravid uterus. 
Hypoventilation can occur following the adminis- 
tration of narcotics for pain relief as well as after a 
period of hyperventilation during painful uterine 
contractions. 41141 

At birth, the fetal circulation undergoes major 
changes that are completed in the first few weeks 
of extrauterine life when the ductus arteriosus clos- 
es permanently. Adequate oxygen tension in the 
blood plays a crucial role in allowing such changes 
to occur. Hypoxemia hinders the normal decrease 
in pulmonary vascular resistance and the closure of 
the ductus arteriosus, resulting in a persistence of 
the fetal circulation. 3 " 7 

However, although hypoxemia must be avoided, 
oxygen should be administered cautiously to neo- 
nates, particularly those born prematurely, because 
a high P a o2 is associated with retinopathy of prema- 
turity, or retrolental fibroplasia. 42 

Hypoxia Secondary to Decreased 
Oxygen Delivery 

Maintenance of aerobic tissue metabolism re- 
quires a minimum end-capillary Po: of about 20 
torr. 43 Below this limit, anaerobic metabolism occurs, 
leading to inadequate utilization of glucose, forma- 
tion of lactic acid, and, eventually, cell death. 3 ' 235 
Decreased oxygen utilization by the peripheral tis- 
sues is observed not only in association with low 
cardiac output but also with distribution of blood 
flow that is abnormal or inappropriate for regional 
needs. Such a situation has been shown to occur in 
patients undergoing valvular heart surgery under 
hypothermic cardiopulmonary bypass. Lactate pro- 
duction and decreased cellular ATP and creatine 
content were observed and were thought to result 
from inadequate oxygen delivery to tissues. The 
persistence of inadequate oxygen delivery in the 
immediate postoperative period led the authors 44 to 
conclude that these patients, who presumably have 
an impaired myocardial function, were unable to 
meet the increase in oxygen demand by increasing 
their cardiac output. This increase in output is nec- 
essary during the postoperative period to meet the 
elevated energy requirements during the hyper- 
metabolic state associated with the stress response 
and spontaneous rewarming. 44 The outcome of crit- 
ical illness, sepsis, and major surgery has been 
shown to be markedly influenced by the patient's 
ability to increase oxygen delivery and, specif- 
ically, cardiac output. 45 ' 46 

Hypoxia Secondary to Decrease in 
Oxidative Metabolism 

Tissue hypoxia is also observed following cya- 
nide poisoning. Exposure to cyanide can be secon- 
dary to inhalation (eg. fumes from domestic fires), 
intentional ingestion (eg, suicide attempts), or iat- 
rogenic administration (eg, sodium nitroprusside 
therapy). Cyanide inhibits cytochrome aa 3 and in- 
terrupts cellular aerobic respiration. 47 Cyanide poi- 
soning is characterized by rapidly developing 
coma, cardiac dysfunction, arrhythmias, and severe 
lactic acidosis secondary to inability of tissues to 
utilize oxygen. In the period preceding cardio- 




respiratory arrest, mixed venous blood has a high 
oxygen saturation, accounting for the absence of 
cyanosis. Substances that bind cyanide are useful in 
the treatment of cyanide intoxication: Sodium thio- 
sulfate combines with cyanide to form thiocy- 
anate. 4s and drugs that induce the conversion of he- 
moglobin to methemoglobin (methylene blue, amyl 
nitrate) lead to the formation of the nontoxic cyan- 
hemoglobin. 44 Because methemoglobin cannot trans- 
port oxygen to the tissues, the use of this mode of 
therapy is limited by the amount of hemoglobin 
that can be safely changed to methemoglobin. 2 

Limits of Tolerance to Hypoxemia 

It is not always possible to decide whether hy- 
poxemia should be treated based on arterial oxygen 
saturation alone. As mentioned earlier, acute desat- 
uration to < 85% is associated with early symptoms 
in normal volunteers, but it is well established that 
adaptation allows many patients to live at sub- 
stantially lower values. In patients with chronic 
pulmonary disease, hypoxemia is often accom- 
panied by hypercapnia and acidosis. A shift of the 
oxygen-hemoglobin dissociation curve to the right 
allows such patients to maintain a normal level of 
consciousness at lower P a o:S than volunteers with 
similar P a <>s breathing at reduced atmospheric 
pressure. In these patients, the level of conscious- 
ness is more dependent on P a co: than on PaO:- 50 Pa- 
tients who have undergone palliative operations for 
cyanotic heart disease with a residual right-to-left 
anatomic shunt are expected to have saturations be- 
low 'normal,' and an arterial saturation > 80% is 
considered acceptable. M However, in the presence 
of a reactive pulmonary vasculature following re- 
pair of a ventricular septal defect or truncus ar- 
teriosus, full oxygen saturation and P a o: > 100 ton- 
may be necessary to reduce pulmonary vascular re- 
sistance and improve right ventricular function." 1 ' 
Special situations such as sickle cell anemia in cri- 
sis mandate not only supplemental oxygen, but ad- 
junctive measures such as hydration, analgesia, and 
transfusion to prevent vaso-occlusive crises. 52 

Ultimately, the lowest acceptable limit of oxy- 
gen saturation is that which prevents organ dam- 
age. Although an hypoxic cerebral injury lasting 
less than 8 minutes has been considered reversible, 3 
irreversible lesions in the hippocampus and Pur- 

kinje cells of the cerebellum have been shown to 
occur after a period of anoxia lasting only 4 min- 
utes. 51 The lower limit for brain survival appears to 
occur at a P a o: of 20-22 torr. 52 

Hypoxemia in Clinical Practice 

Desaturation in Special Clinical Settings 

The surgical patient, particularly during general 
anesthesia, is at high risk for developing hypox- 
emia. An increase in alveolar-to-arterial-oxygen- 
tension gradient occurs after premedication and af- 
ter induction of anesthesia, particularly if inter- 
mittent positive pressure ventilation is employed. 54 
This effect is more pronounced with age and may 
also extend into the postoperative period. Although 
atelectasis occurs and can be important following 
intrathoracic procedures, this is not now thought to 
be the most common cause of anesthesia-related 
hypoxemia. A fall in cardiac output may cause in- 
creased tissue oxygen extraction, resulting in a de- 
crease in mixed-venous oxygen content, which in 
the presence of ventilation-perfusion mismatch 
lowers arterial oxygen tension. Alterations of ven- 
tilation-perfusion relationships secondary to pos- 
itive pressure ventilation, decreased functional re- 
sidual capacity, and possible blunting of hypoxic 
pulmonary vasoconstriction (HPV) can result in hy- 
poxemia. Attempts to substantiate the effect of an- 
esthetics on HPV have yielded conflicting results. 
The inhibitory effect of inhaled anesthetics varies, 
depending on the experimental setting. Sykes and 
colleagues demonstrated abolition of HPV in iso- 
lated perfused cat lung during administration of 
halothane, ether, and nitrous oxide." 1 "' In dogs, in- 
hibition of HPV was found to be dose related and 
significant with nitrous oxide, isoflurane, and flu- 
roxene but not with halothane and enflurane." 1 ' 1 " 17 
However, in patients undergoing thoracic surgery 
with single-lung ventilation in the lateral decubitus 
position, halothane and isoflurane do not sig- 
nificantly inhibit HPV. 58 ' 59 Factors such as pul- 
monary artery pressure, patient's position, the pres- 
ence of inflammation that increases blood flow or 
of disease that limits blood flow in the non- 
ventilated lung play a role in determining how 
much blood will flow in the nonventilated area. h " In 
contrast to volatile anesthetics, intravenous an- 
esthetics do not interfere with HPV. 18 




Because of the known decrease in P a o: associat- 
ed with general anesthesia, it is nearly universal for 
anesthesiologists and anesthetists to provide higher 
levels of oxygen during both surgery and the early 
recovery period. Diffusion hypoxemia occurs tran- 
siently after nitrous oxide anesthesia. If the patient 
is breathing room air following an oxygen-nitrous 
oxide anesthetic, nitrogen diffuses into the blood 
while nitrous oxide diffuses out of the blood into 
the alveoli (Fig. 2). Due to solubility differences, 
the volume of nitrous oxide exiting the blood ex- 
ceeds the volume of nitrogen entering the circula- 
tion, thus diluting the residual oxygen in the alveo- 
li. 61 Provision of supplemental oxygen during the 
elimination of nitrous oxide helps avoid hypox- 
emia. Airway obstruction, laryngospasm, and de- 
pressed ventilation due to residual opioids or neu- 
romuscular blocking drugs can also contribute to 
hypoxemia in the early postoperative period. 

< N 2 +0 2 

CO2 + N2O 

Fig. 2. Solubility differences between nitrous oxide 
(N 2 0) and nitrogen (N 2 ) lead to diffusion hypoxemia in 
patients breathing room air following nitrous oxide an- 

A depression of ventilatory response that results 
in apnea and/or hypoxemia occurs to an alarming 
extent following opioid or benzodiazepine sedation 
for 'minor' procedures such as bronchoscopy and 
gastrointestinal endoscopy. Hypoxemia, defined as 
an oxyhemoglobin saturation < 90%, occurs in half 
of normal volunteers given a sedative dose of fen- 
tanyl (2 pg/kg). The combination of fentanyl and 
midazolam caused hypoxemia in 11 of 12 subjects 
and apnea in 6 of 12. 62 One study of dental out- 
patients given nitrous oxide-oxygen-halothane an- 
esthesia (before the era of pulse oximetry) doc- 

umented a 20% incidence of hypoxemia caused by 
respiratory obstruction unrecognized by the an- 
esthetist. 63 Hypoxemia during upper gastroin- 
testinal endoscopy can be minimized by titrating 
small amounts of sedation until the desired effect is 
obtained rather than administering a standard (usu- 
ally larger) bolus. 64 In another study, provision of 
supplemental oxygen by the oral rather than nasal 
route was effective in diminishing the incidence of 
hypoxemia. to These patients were noted to breathe 
through the mouth while the esophagus was in- 

In view of the previous observations in relatively 
healthy individuals, it is not surprising to find that 
hypoxemia is common in patients undergoing car- 
diac catheterization. Monitoring of saturation via 
pulse oximetry (S p r>) in these circumstances re- 
vealed that all patients had at least one episode of 
desaturation below 90%. The number of such epi- 
sodes was highest in patients with severe impair- 
ment of left ventricular function. 66 Hypoxemia is of 
particular importance in such patients because their 
underlying disease limits oxygen delivery to the 
myocardium and peripheral tissues, and hypoxemia 
exacerbates a situation that is already critical. 

The preceding data highlight the importance of 
monitoring oxygenation in patients receiving seda- 
tion to facilitate diagnostic procedures. Administra- 
tion of supplemental oxygen is often necessary, and 
medical personnel skilled in airway management 
should be available, particularly when opioids and 
benzodiazepines are given in combination. 

Reliability of Signs and Symptoms 

Until the advent of pulse oximetry, clinical de- 
tection of hypoxemia could be assured only with 
arterial blood gas analysis. Although the onset of 
new neurologic signs suggestive of hypoxemia 
might direct the clinician to the appropriate diag- 
nosis, such observations may be unreliable or im- 
possible to elicit in the patient who is critically ill 
or under general anesthesia. Because the brain is 
the organ most sensitive to oxygen deprivation, 
substantial hypoxic neurologic damage may occur 
before cardiac arrhythmias and organ system dys- 
function become apparent. 

Most clinicians have been trained to look for 
cyanosis, or bluish discoloration of the skin and 




mucous membranes, despite the fact that cyanosis 
is a relatively unreliable indicator of hypoxemia. 
Furthermore, even experienced observers do not 
reliably detect cyanosis even when oxygen satura- 
tion is as low as 70-80%. M Cyanosis occurs when 
the concentration of deoxygenated hemoglobin is 
sufficient to produce the characteristic color chang- 
es. It reflects the presence of an absolute concentra- 
tion of desaturated hemoglobin in the blood rather 
than the overall oxygen content. About 5 g/dL of 
deoxyhemoglobin must be present in the blood to 
produce cyanosis. 

In the normal individual with a hemoglobin of 
15 g/dL, full saturation of the arterial blood results 
in an oxygen-carrying capacity of 20 mL/dL of 
blood. If we assume a normal tissue extraction of 5 
vol% of oxygen, the end-capillary blood contains 
only 4.1 mg/dL of reduced hemoglobin. If a pul- 
monary process results in hypoxemia, the arterial 
blood will contain some deoxyhemoglobin to 
which the deoxyhemoglobin produced by normal 
tissue extraction is added. The resultant concentra- 
tion of deoxyhemoglobin in the venous blood may 
then be > 5 g/dL, causing cyanosis to appear. In the 
severely anemic individual with a hemoglobin of < 
5 g/dL. even complete desaturation of the blood 
does not produce a concentration of deoxyhemo- 
globin > 5 g/dL, and cyanosis does not occur. Con- 
versely, a polycythemic individual with mild arteri- 
al desaturation may demonstrate cyanosis even in 
the absence of oxygen lack, particularly when ex- 
posed to cold or in the presence of slow blood flow 
in the extremities. In such individuals, even a nor- 
mal oxygen extraction results in a concentration of 
deoxyhemoglobin > 5 g/dL. 2 

Other situations in which reduced arterial oxy- 
gen content (but normal P a o:) can be present with- 
out cyanosis include carbon monoxide poisoning 
and metabolic abnormalities preventing oxygen 
utilization by the cells (eg, cyanide poisoning). 
Carbon monoxide poisoning produces a severe, 
acute functional anemia by reacting directly with 
hemoglobin to form carboxyhemoglobin. Carboxy- 
hemoglobin cannot release oxygen and causes a 
shift of the dissociation curve of the remaining nor- 
mal hemoglobin to the left, decreasing oxygen re- 
lease to the tissues even further. Carbon monoxide 
has an affinity for hemoglobin 210 times that of ox- 
ygen, and it has an elimination half-life of about 4 

hours assuming normal ventilation.' :1 ^"Treatment 
of carbon monoxide poisoning relies on the ad- 
ministration of pure or hyperbaric oxygen to in- 
crease the amount of oxygen dissolved in plasma 
and facilitate the displacement of carbon monoxide 
from the hemoglobin molecule. 

As the use of continuous pulse oximetry moni- 
toring has become more widespread, its merits 
have become more obvious. Analysis of anesthetic- 
related closed malpractice claims filed between 
1974 and 1988 suggests that application of pulse 
oximetry plus capnometry can identify 93% of pre- 
ventable anesthetic mishaps. 6 * Intraoperative mon- 
itoring of pulse oximetry in children has been 
shown to increase safety by allowing earlier de- 
tection and correction of episodes of hypoxemia. 69 

Reliability of Oximeter Alarms 

Many events including hypothermia and hypo- 
perfusion can cause a pulse oximeter to falsely in- 
dicate a low arterial oxygen saturation. However, 
such information should never be discarded with- 
out further investigation because of the serious con- 
sequences of an unrecognized episode of hypoxia. 
There is no doubt that oximetry provides earlier 
warning of adverse events related to failure of oxy- 
genation or ventilation in the operating room. Fac- 
tors influencing the accuracy and reliability of 
pulse oximeters as well as the incidence of false 
alarms with various monitoring techniques and de- 
vices are discussed elsewhere in this journal. 70 

Differential Diagnosis of Hypoxia 
and Its Management 

Assuming that hypoxia is truly present, the ap- 
proach to the differential diagnosis can be made ac- 
cording to the scheme presented in Figure 3. Sever- 
al aspects of management are addressed in detail 
elsewhere in this issue and are presented here brief- 
ly. The eight major reasons for failure of tissue ox- 
ygenation correspond to those in Figure 1 . Workup 
should start by measurement of the inspired oxygen 
concentration and evaluation of its appropriateness 
to the environmental conditions (eg. high altitude, 
single-lung ventilation). Then, administration of 
supplemental oxygen should be implemented if in- 
dicated. Assessment of ventilation should include 





4 and 5- 

Hearl Catheter 
and/or Echo 

1. LowF|o 2 

2 Ventilation 

3 Oxygenation 

4 Venous-Arterial 


5 Circulation 
6. Hemoglobin 

7 Consumption 

8 Utilization 

8. Blood Work 

Fig. 3. Schema for differential diagnosis of hypoxia and 

examination and auscultation of the chest, spirom- 
etry, and arterial or end-tidal Pco2- When an in- 
tubated patient develops signs of hypoxemia in the 
operating room or intensive care unit, it is essential 
to manually ventilate him with 100% oxygen from 
a self-inflating bag while an assistant examines the 
oxygen supply circuit. Immediate improvement of 
arterial oxygen saturation using this alternative 
ventilatory mode quickly identifies equipment mal- 
function as the cause of the problem. However, if 
no improvement occurs, auscultation of the chest 
for detection of bronchospasm and passage of a 
suction catheter to exclude kinking or occlusion of 
the endotracheal tube can be most helpful, par- 
ticularly in pediatric patients in whom inspissated 
secretions may easily clog a small-bore airway. 
Monitoring of end-tidal C0 2 eliminates dis- 
lodgment or misplacement of the endotracheal tube 
as a cause of concern. 

Assessment of adequacy of arterial oxygenation 
must start by measuring arterial oxygen saturation. 
Evaluation of arterial hypoxemia includes clinical 
examination, a chest radiograph, and a search for 
causes of venoarterial shunts, including ventilation- 
perfusion mismatch (eg, atelectasis, pneumonia, 
pulmonary embolus), impaired oxygen diffusion 
across the alveolocapillary membrane (eg, pulmo- 
nary edema, interstitial pneumonitis), and presence 
of anatomic shunts (eg, cyanotic heart disease or an 
acquired ventriculoseptal defect). 

Assessment of oxygen delivery to tissues and or- 
gans requires global as well as regional measure- 

ment of blood flow and oxygen transport. Clinical 
examination includes measurement of blood pres- 
sure, peripheral pulses, and skin temperature and 
color. More information can be obtained by direct 
measurement of cardiac output using thermodilu- 
tion, dye-dilution, or echocardiography. Measure- 
ment of hemoglobin concentration and specialized 
oximetry and hemoglobin electrophoresis provides 
information about the oxygen-carrying capacity of 
arterial blood. Common causes of acute circulatory 
failure include congestive heart failure, cardiac ar- 
rest, ventricular dysrhythmias, large pulmonary or 
peripheral emboli, tension pneumothorax, cardiac 
tamponade, acute valvular dysfunction, aortic dis- 
section, and, rarely, severe hypocalcemia. Man- 
agement according to Advanced Cardiac Life Sup- 
port protocol with specific actions to correct the 
remediable causes (eg, relieving tamponade and 
tension pneumothorax) addresses this problem. 
Some tertiary centers may be able to intervene in 
acute embolic and cardiac events, if it is possible to 
place the patient on cardiopulmonary bypass im- 
mediately. Even with such care the mortality is 
likely to be high. 

Compared to global oxygen transport, regional 
oxygen delivery is more difficult to assess. Many 
monitoring and diagnostic tools provide indirect in- 
formation about the adequacy of organ perfusion 
by evaluating organ function. Such tools include 
electrocardiography, hemodynamic monitoring, echo- 
cardiography, electroencephalography, and biochem- 
ical indicators of hepatic, renal, and pancreatic 
functions (lactic acidosis, liver function tests, ser- 
um creatinine, amylase and lipase). Unfortunately, 
biochemical indicators are slow to reveal an hypox- 
ic insult and, therefore, corrective actions based on 
them are often delayed. 

The last, though not the least, important step in 
the differential diagnosis of hypoxia consists of 
looking at the balance between tissue oxygen needs 
and oxygen supply as well as the ability of cells to 
utilize the available oxygen. Increased oxygen con- 
sumption occurs in many situations, including fe- 
ver, malignant hyperthermia, sepsis, hyperthyroid- 
ism, and neuroendocrine response to stress. Meta- 
bolic acidosis in the absence of low cardiac output 
and arterial and venous desaturation is usually a 
sign of inhibition of oxidative phosphorylation by 
cyanide, particularly when a history of exposure to 




fumes, therapy with sodium nitroprusside, or sui- 
cide attempt can be elicited. 

Summary and Conclusions 

Although moderately severe tissue hypoxia has 
important adverse effects on neurologic, cardiac, 
pulmonary, renal, and gastroenteric functions, se- 
vere hypoxia is devastating. Unfortunately, many 
of the clinical signs associated with hypoxia be- 
come apparent only after irreversible neurologic 
damage has occurred. Because no single, clinically 
available, monitoring technique is both sensitive 
and specific enough to provide an immediate and 
reliable warning of hypoxemia and hypoxia, it is 
essential to observe patients closely and use multi- 
ple modalities of monitoring, including pulse oxim- 
etry, to ensure early detection and correction of 
such problems. 


1. Stedman TL. Stedman's medical dictionary. 25th ed. 
Baltimore: Williams & Wilkins, 1990:756. 

2. Youmans WB. Hypoxia, asphyxia, dysbarism, oxygen 
therapy, and resuscitation. In: Brobeck JR, ed. Best & 
Taylor's physiological basics of medical practice, 9th 
ed. Baltimore: Williams & Wilkins, 1973:63-90. 

3. Goudsouzian N. Karamanian A. Physiology for the an- 
esthesiologist. 2nd ed. Norwalk CT: Appleton-Century- 
Crofts, 1984. 

4. Dickinson JG. High altitude cerebral edema: cerebral 
acute mountain sickness. Semin Respir Infect 1983:5: 

5. York EL, Jones RL, Menon D, Sproule BJ. Effects of 
secondary polycythemia on cerebral blood flow in 
chronic obstructive pulmonary disease. Am Rev Respir 
Dis 1980;121:813-818. 

6. Brenner RP. The electroencephalogram in altered states 
of consciousness. Neurol Clin 1985:3(3):6 15-631. 

7. Dulfano MJ, Ishikawa S. Hypercapnia: mental changes 
and extrapulmonary complications. Ann Intern Med 

8. Phillipson EA. Regulation of breathing during sleep. 
Am Rev Respir Dis 1977;1 15(6, Part 2):217-224. 

9. Guilleminault C. Tilkian A, Dement WC. The sleep ap- 
nea syndromes. Annu Rev Med 1976:27:465-484. 

10. Rolfe I, Olson LG, Saunders NA. Long-term acceptance 
of continuous positive airway pressure in obstructive 
sleep apnea. Am Rev Respir Dis 1991;144(5):1 130- 

1 1 . Fletcher EC, Lesske J, Qian W, Miller CC III, Unger T. 
Repetitive, episodic hypoxia causes diurnal elevation of 
blood pressure in rats. Hypertension 1992:19(6, Part 

12. Van Den Arrdweg JG. Karemaker JM. Repetitive ap- 
neas induce periodic hypertension in normal subjects 
through hypoxia. J Appl Physiol 1992;72(3):821-827. 

13. McDonald DM. Role of glomus cells as dopaminergic 
interneurons in the chemoreceptive function of the ca- 
rotid body. Adv Biochem Psychopharmacol 1977:16: 

14. Kressin NA. Nielsen AM. Laravuso R. Bisgard GE. 
Domperidone-induced potentiation of ventilatory re- 
sponses in awake goats. Respir Physiol 1986:65(2): 169- 

15. Walley KR, Becker CJ, Hogan RA, Teplinsky K, Wood 
LD. Progressive hypoxemia limits left ventricular oxy- 
gen consumption and contractility. Semin Anesthesia 

16. Heavner JE, Dryden CF Jr. Sanghani V. Huemer G, 
Bessire A, Badgwell JM. Severe hypoxia enhances cen- 
tral nervous system and cardiovascular toxicity of bupi- 
vacaine in lightly anesthetized pigs. Anesthesiology 

17. Von Euler US. Liljestrand G. Observations of the pul- 
monary artery blood pressure in the cat. Acta Physiol 
Scand 1946:12:301-307. 

18. Bergman NA. Hypoxic pulmonary vasoconstriction. 
Semin Anesthesia 1 987;6( 3 ): 1 88- 1 95 . 

19. Sobol BJ, Bottex G, Emirgil C, et al. Gaseous diffusion 
from alveoli to pulmonary vessels of considerable size. 
CircRes 1963:13:71. 

20. Staub NC. Site of hypoxic pulmonary vasoconstriction. 
Chest 1985;88(4, Suppl):240S-245S. 

21. Hales CA. The site and mechanism of oxygen sensing 
for the pulmonary vessels. Chest 1985:88(4. Suppl): 

22. Lloyd TC Jr. Influences of Po: and pH on resting and 
active tensions of pulmonary arterial strips. J Appl 
Physiol 1967;22:1101-1109. 

23. Barer GR. Reactivity of the vessels of collapsed and 
ventilated lungs to drugs and hypoxia. Circ Res 1966; 

24. Hauge A. Role of histamine in hypoxic pulmonary hy- 
pertension in the rat: I. Blockade or potentiation of en- 
dogenous amines, kinins and ATP. Circ Res 1968:22: 

25. Radermacher P. Santak B, Becker H, Falke KJ. Pros- 
taglandin Ei and nitroglycerin reduce pulmonary cap- 
illary pressure but worsen ventilation-perfusion dis- 
tributions in patients with adult respiratory distress 
syndrome. Anesthesiology 1989:70:601-606. 

26. Bakhle YS, Ferreira SH. Lung metabolism of eicosan- 
oids: prostaglandins, prostacyclin, thromboxane and leu- 
kotrienes. In: Handbook of physiology. Washington 
DC: American Physiological Society, 1985:3( 1 ):365. 

27. Palmer RMJ. Ferrige AG. Moncada S. Nitric oxide re- 
lease accounts for the biological activity of endotheli- 
um-derived relaxing factor. Nature 1987:327:524-526. 

28. Burghuber OC. Nifedipine attenuates acute hypoxic pul- 
monary vasoconstriction in patients with chronic ob- 
structive pulmonary disease. Respiration 1987:52:86-93. 




29. Benumof JL. Hypoxic pulmonary vasoconstriction and 
infusion of sodium nitroprusside (editorial). Anesthesi- 
ology 1979;50:481-483. 

30. Prewitt RM, Wood LDH. Nitroprusside increases pul- 
monary shunt and improves left ventricular function in 
dogs with oleic acid pulmonary edema (abstract). Am 
Rev Respir Dis 1978:1 17(4. Part 2):382. 

31. Simonneau G, Escourrou P. Duroux P, Lockhart A. In- 
hibition of hypoxic pulmonary vasoconstriction by ni- 
fedipine. N Engl J Med 198 1 ;304(26): 1582-1585. 

32. Furman WR. Summer WR, Kennedy TP. Sylvester JT. 
Comparison of the effects of dobutamine, dopamine, 
and isoproterenol on hypoxic pulmonary vasocon- 
striction in the pig. Crit Care Med 1982; 10(6):37 1-374. 

33. Epstein RM. Wheeler HO. Frumin M. Habif DV. Pap- 
per EM, Bradley SE. The effect of hypercapnia on es- 
timated hepatic blood flow, circulating splanchnic blood 
volume, and hepatic sufobromophthalein clearance dur- 
ing general anesthesia in man. J Clin Invest 1961; 

34. Miller TA. Mechanisms of stress-related mucosal dam- 
age. Am J Med 1987;83(6A):8-14. 

35. Fiddian-Green RD. McGough E. Pittenger G, Rothman 
E. Predictive value of intramural pH and other risk fac- 
tors for massive bleeding from stress ulceration. Gastro- 
enterology 1983;85:613-620. 

36. Aranha GV, Pickleman J. Pifarre R, Scanlon PJ. Gunnar 
RM. The reasons for gastrointestinal consultation after 
cardiac surgery. Am Surg 1984;50:301-304. 

37. Haas GS, Warshaw AL, Daggett WM. Aretz HT. Acute 
pancreatitis after cardiopulmonary bypass. Am J Surg 

38. Fiddian-Green RG, Amelin PM, Herrmann JB, Arous E, 
Cutler BS, Schiedler M. et al. Prediction of the develop- 
ment of sigmoid ischemia on the day of aortic opera- 
tions: indirect measurements of intramural pH in the co- 
lon. Arch Surg 1986;121:654-660. 

39. Gordon AS. Zanjani ED. Hormones and erythropoiesis. 
In: Greenwalt Tl, Jamieson GA, eds. Formation and de- 
struction of blood cells. Philadelphia: JB Lippincott. 

40. Cheek TG. Gutsche BB. Maternal physiologic altera- 
tions during pregnancy. In: Schneider SM. Levinson G, 
eds. Anesthesia for obstetrics, 2nd ed. Baltimore: Wil- 
liams & Wilkins, 1987:3-13. 

41. Hueh A, Huch R, Schneider H, Rooth G. Continuous 
transcutaneous monitoring of fetal oxygen tension dur- 
ing labour. Br J Obstet Gynaecol 1977;84(Suppl 1):1- 

42. Benumof JL. Respiratory physiology and respiratory 
function during anesthesia. In: Miller RD. ed. Anesthe- 
sia, 3rd ed. Vol 1. New York: Churchill Livingstone. 

43. Landis EM, Pappenheimer JR. Exchange of substances 
through the capillary walls. In: Hamilton WE. ed. Hand- 
book of physiology: circulation. Vol II, Chapter 29. 
Washington DC: American Physiological Society. 1963. 

44. Fiaccadori E, Vezzani A, Coffrini E, Guariglia A, Ron- 
da N, Tortorella G, et al. Cell metabolism in patients un- 
dergoing major valvular heart surgery: relationship with 

intra and postoperative hemodynamics, oxygen trans- 
port, and oxygen utilization patterns. Crit Care Med 

45. Shoemaker WC, Appel PL. Kram HB, Waxman K. Lee 
T-S. Prospective trial of supranormal values of sur- 
vivors as therapeutic goals in high-risk surgical patients. 
Chest 1988;94:1176-1186. 

46. Shoemaker WC, Appel PL. Kram HB. Tissue oxygen 
debt as a determinant of lethal and nonlethal post- 
operative organ failure. Crit Care Med 1988:16(11): 

47. Taitelman U. Acute cyanide poisoning. In: Hall JB, 
Schmidt GA. Wood LDH. eds. Principles of critical 
care. New York: McGraw-Hill, 1992:2125-2129. 

48. Chen KK, Rose CL, Clowes GHA. Comparative values 
of several antidotes in cyanide poisoning. Am J Med Sci 

49. Hug EL. Intoxication par acid cyanhydrique. C R Soc 
Biol (Paris) 1933;112:511-513. 

50. Refsum HE. Relationship between state of conscious- 
ness and arterial hypoxemia and hypercapnia in patients 
with pulmonary insufficiency, breathing air. Clin Sci 

51. Nikaidoh H. Leonard SR. Morris-Copeland M. Car- 
diothoracic surgery: perioperative principles. In: Levin 
DL. Morriss FC, eds. Essentials of pediatric intensive 
care. St Louis: Quality Medical Publishing Inc. 1990: 

52. Sanders JM. Life-threatening complications of sickle 
cell disease. In: Levin DL, Morriss FC, eds. Essentials 
of pediatric intensive care. St Louis: Quality Medical 
Publishing Inc. 1990:383-385. 

53. Weinberger LM. Gibbon MH, Gibbon JH Jr. Temporary 
arrest of the circulation to the central nervous system. 
Arch Neurol Psych 1940;43(4):615-634. 

54. Leigh JM. The pulmonary circulation. In: Scientific 
foundations of anaesthesia — the basics of intensive care, 
4th ed. Chicago: Oxford/Year Book Medical Publishers. 

55. Sykes MK. Davies DM. Chakrabarti MK. Loll L. The 
effect of inhalation anaesthetic agents on the pulmonary 
vasculature of the isolated perfused cat lung. Br J An- 
aesth 1973:45:114-115. 

56. Domino KB. Borowec L, Alexander CM. Williams JJ. 
Chen L, Marshall C. et al. Influence of isoflurane on hy- 
poxic pulmonary vasoconstriction in dogs. Anesthes- 
iology 1986;64:423-429. 

57. Mathers J, Benumof JL, Wahrenbrock EA. General an- 
esthetics and regional hypoxic pulmonary vasocon- 
striction. Anesthesiology 1977:46:1 11-114. 

58. Rogers SN. Benumof JL. Halothane and isoflurane do 
not impair arterial oxygenation during one lung ventila- 
tion in patients undergoing thoracotomy (abstract). 
Anesthesiology 1983:59:A532. 

59. Augustine SD, Benumof JL. Halothane and isoflurane 
do not impair arterial oxygenation during one lung ven- 
tilation in patients undergoing thoracotomy (abstract). 
Anesthesiology 1984;61:A484. 

60. Benumof JL. One-lung ventilation and hypoxic pul- 
monary vasoconstriction: implications lor anesthetic 
management. Anesth Analg 1985;64:821-833. 







Dripps RD. Eckenhoff JE, Vandam LD. The immediate 
postoperative period: recovery and intensive care. In: In- 
troduction to anesthesia: the principles of safe practice. 
6th ed. Philadelphia: WB Saunders Co, 1982:410-420. 
Bailey PL. Pace NL, Ashburn MA, Moll JW, East KA, 
Stanley TH. Frequent hypoxemia and apnea after seda- 
tion with midazolam and fentanyl. Anesthesiology 

Al-Khishali T, Padfield A, Perks ER, Thornton JA. Car- 
diorespiratory effects of nitrous oxide: oxygemhalo- 
thane anaesthesia administered to dental outpatients in 
the upright position. Anaesthesia 1978:33(2): 1 84- 1 88. 
Bell GD, Antrobus JH, Lee J, Coady T. Morden A. Bo- 
lus or slow titrated injection of midazolam prior to 
upper gastrointestinal endoscopy? Relative effect on ox- 
ygen saturation and prophylactic value of supplemental 
oxygen. Aliment Pharmacol Ther 1990;4(4):393-401. 
Bell GD, Antrobus JH, Lee J, Coady TJ, Morden A. Pat- 
tern of breathing during upper gastrointestinal endos- 

copy: implications for administration of supplemental 
oxygen. Aliment Pharmacol Ther 1991:5(4):399-404. 

66. Dodson SR, Hensley FA Jr. Martin DE, Larach DR. 
Morris DL. Continuous oxygen saturation monitoring 
during cardiac catheterization in adults. Chest 1988:94 

67. Comroe JH Jr. Botelho S. The unreliability of cyanosis 
in the recognition of arterial anoxemia. Am J Med Sci 

68. Tinker JH. Dull DL, Caplan RA, Ward RJ, Cheney FW. 
Role of monitoring devices in prevention of anesthetic 
mishaps: a closed claims analysis. Anesthesiology 1989; 

69. Cote CJ, Goldstein EA. Cote MA, Hoaglin DC, Ryan 
JF. A single-blind study of pulse oximetry in children. 
Anesthesiology 1988:68(2): 1 84- 1 88. 

70. Hess D, Kacmarek RM. Techniques and devices for 
monitoring oxygenation. RespirCare 1993:38:646-671. 

Higgins Discussion 

Hess: Let's open the floor to ques- 
tions. And, let me remind you, before 
we get into this, that the focus of this 
morning's discussion is patho- 
physiology. I think it could be easy 
for us now to launch into a discus- 
sion of treatment and management 
styles, but 1 think we could probably 
save some of the discussion regard- 
ing various treatment approaches un- 
til tomorrow afternoon. 

Pierson: It strikes me that many of 
the signs and other manifestations 
that were described for the presence 
of hypoxia are also present in critical 
illness in general, which emphasizes 
the need for things other than were 
described as clinical observations to 
assess the clinical importance of hy- 
poxia. I wonder how people respond 
to that. 

Hess: Your point is that some of the 
signs of hypoxia are neither sensitive 
nor specific . . . 

Pierson: Right, and if we're going to 
accurately assess hypoxia and know 
whether it is harmful for our patient, 
we're going to need other tools be- 
sides just the clinical observations. 

Hess: First we need to know if it's 

Huang: I wonder about Dr Pierson' s 
or Frank's (Cerra) viewpoint on how 
to assess tissue oxygenation. We all 
agree that the present method we 
have to evaluate tissue oxygenation 
in clinical practice is inadequate. We 
refer to tissue hypoxia by looking at 
something that is coming into the tis- 
sue without knowing what's going 
on in the tissue. I think more effort 
should be put into studying how the 
tissues use oxygen, especially under 
pathological conditions. For exam- 
ple, the mitochondrial electron-trans- 
port chain consumes most of the ox- 
ygen delivered to the cell, and some 
methods have been developed in re- 
cent years to assess this process. I 
don't know how useful that will be 
clinically, but I think the approach is 
in the right direction. 

Cerra: My comments relate to tissue 
hypoxia and the common surrogates 
that we use in laboratory tests for the 
measurement of organ function and 
the clinical analysis of central ner- 
vous system function (ie, encepha- 
lopathies). One can show, clinically 
and experimentally, the induction and 
progression of encephalopathy and 
organ dysfunctions in the absence of 
any of the usual criteria of tissue hy- 
poxia that we use, including this con- 
cept of flow-dependency. I think this 

creates a real problem for us in terms 
of how we're going to sort out this 

Wood: I agree with all those com- 
ments, and I just want to rephrase Dr 
Higgins' paper in the way that I see 
his target here. In the context of Dr 
Pierson' s paper and the discussion, 
we are interested in tissue oxy- 
genation, and one part of it starts with 
the detection of arterial hypoxia. The 
question can be asked concerning 
critically ill patients. How low can 
you go? How low is the arterial O: 
saturation that is associated with end- 
organ dysfunction? Now. that's a pe- 
culiar question, which is asked only 
when there's a problem on the other 
side — for example, with the inspired 
oxygen at 100% and the arterial sat- 
uration at 80%. Then some tricks are 
employed to improve the oxygenation 
by decreasing the shunt. The best one 
is PEEP, but that drops the cardiac 
output. At the limit, a conundrum 
(which asks the question that comes 
up in Dr Higgins' paper) is. How low 
can you allow arterial saturation to go 
before causing harm? It is surprising 
to me how few studies have actually 
looked at that. 

Dr Pierson showed the oxyhemo- 
globin dissociation curve and pointed 
out that we all get 'antsy' when we 




start down the slippery slope. At 
90% saturation or less you lose a lot 
of content for every fall in Po> there- 
fore, the saturation meter is really 
good at detecting 90% — well, wait a 
minute! 90% might require Fio; 1.0; 
80% saturation might require only 
F102 0.70 in the same patient. How 
low you can go and what the limits 
of S a O: are on myocardial function 
has been studied. Keith Walley 1 
measured the coronary blood flow 
and ventricular mechanics as he pro- 
gressively dropped the arterial sat- 
uration in anesthetized dogs. At 
about 75% SaO:- the end-systolic left- 
ventricular volume began to in- 
crease, the heart rate slowed, and 
there was a sudden, irreversible car- 
diac arrest! I see that sequence often 
enough in patients who are on the 
edge of hypoxia that I don't want to 
approach S a o : of 75%. In that study, 
myocardial metabolism was main- 
tained as Sao ; decreased because cor- 
onary blood flow progressively in- 
creased until it couldn't increase 
anymore — massive coronary va- 
sodilation and any further reduction 
in the arterial content was now as- 
sociated with anaerobic metabolism 
of the heart. I think that's a paradigm 
for every organ dysfunction. 

Dr Cerra talked about the central 
nervous system. Most of the patients 
who have the clinical scenario that I 
just talked about are unconscious or 
have been rendered unconscious by 
medications that help their critical 
therapy. In those circumstances, I so 
often lose (and wish I didn't) some 
handle on what the CNS function is. 
Now, how low you can go isn't just 
the saturation anymore, it's what's 
the cerebral blood flow? What's the 
level of cerebral edema that is mak- 
ing a critical intracranial vascular 
waterfall and limiting blood flow? 
And often we don't have a handle on 
that. That's more important, I think, 
than How low is the saturation? If 
that's the heart and the CNS limits, 
what about the kidney? The kidney 
sits right on the edge of hypoxia, 
normally. The Po : at the tip of the 

loop of Henle is something like 25 
torr before counter-current diffusion 
increases Po; toward 50 torr as blood 
comes back up toward the renal cor- 
tex. If you lower the arterial satura- 
tion a little bit, you end up with pre- 
renal oliguria. So, I think that's what 
I hear Dr Higgins doing. Here the 
saturation shouldn't go below 80%. 
That's where to set the alarm. Well, 
wait a minute, that's too low for 
most people. And when you set it 
low, you better have some handle on 
what the other contributors are to ox- 
ygen delivery to the tissues. Lastly, 
the metabolic rate of the tissues at 
that time influences the Do: at which 
hypoxia occurs. 

1. Walley KR. Becker CJ, Hogan RA. 
Teplinsky K. Wood LDH. Progres- 
sive hypoxemia limits left ventricular 
oxygen consumption and contractil- 
ity. Circ Res 1988;63:849-859. 

Hess: I think a practical sort of bed- 
side corollary question might be, 
How low can what go? — your arteri- 
al saturation, your arterial Po> mixed 
venous saturation, the oxygen con- 
sumption, the oxygen delivery. 

Nelson: We might as well start hav- 
ing some controversy. I don't think 
the question in any way is. How low 
can you go? I think that is an ab- 
solutely moot point, and a more im- 
portant question is some product of 
severity and duration that should 
concern us as clinicians. Clearly, it 
is. How long can it be low? Because 
all of us who have seen patients in 
profound shock for a period of time 
recognize that they can be bad (what 
they call in Nashville "low sick") for 
a few minutes, have the worst set of 
metabolic parameters and oxygena- 
tion parameters you've ever seen, 
and then bounce back after resus- 
citation. They'll be fine. They have a 
very good outcome. The patient who 
comes to us after the motor vehicle 
crash 6 hours away, poor resus- 
citation, treated outside a hospital, in 
shock for 6 or 8 hours by the time we 
see him, has a terrible course. He 

may not have been as hypoxic, he 
may not have been as acidemic as 
the other patient, but the slow re- 
suscitation. I think, has a lot more 
impact than just how low it has gone. 

Cerra: I would like to expand this a 
little more because I'm not sure we 
have the right criteria on the table 
now. The issue here is. How do you 
know (if you're sitting on the cell) 
that you have delivery equal to de- 
mand? In any study in which we 
can't measure demand, a functional 
end point, to define delivery end 
point parameters categorically is con- 
founding. We also have the problem 
of measuring total body metabolism 
and individual organ metabolism. 
That's where I get stuck. I think the 
issue is. How do you know when de- 
livery is equal to demand? You can 
manipulate and modulate all of that 
list of oxygenation parameters; how- 
ever, as long as you achieve the end 
point itself, these parameters don't 
seem to make any difference. 

Conway: I suppose the question be- 
comes. How do you know when 
you've achieved an end point, if you 
really aren't able to tell definitively 
that you've met the demand at the 
tissue level? This concept — How 
low can you go? — is really mak- 
ing the rounds of this table. Why are 
we so concerned with this? Part of 
the attraction is the concept of do no 
harm. Some of us are uncomfortable 
with our patients on 70% inspired 
oxygen, whereas others are uncom- 
fortable at 40%. We're focused on 
keeping our intervention level as low 
as we can. and letting the resulting 
values go as low as we can, to avoid 
exposing the patient to harm from 
the intervention. We also have a lot 
of people talking to us about the cost 
of unjustified care, and why are we 
doing this intervention without a 
cost-benefit analysis, and so on. This 
whole discussion, and the one that 
preceded it, points out clearly that 
we are dealing with clinical evalua- 
tion tools that aren't up to the task of 
telling us what we arc and are not 




achieving for our patients. We know 
that many of the signs and symptoms 
are nonspecific and inadequate for 
the task. We don't know what's hap- 
pening at the tissue level, and we 
never have. We've known for a long 
time that P a o: is not the actual an- 
swer, that we have to look at content 
and delivery issues. But we generally 
have become too comfortable and 
too good at what I call intuitive treat- 
ment. We need to move away from 
intuitive treatment. We need to come 
up with a mechanism that reminds us 
when other pieces of information 
haven't been assessed yet. We need 
to work toward a means of assim- 
ilating all the information we have, 
to work toward pulling all the in- 
formation available into a concise 
whole that tells us something of sig- 

Hudson: I don't think you're ever 
going to be able to do that. You 
could put all these things into an al- 
gorithm and come out with absolute- 
ly the wrong answer. It's not to say I 
don't think we should be trying to 
get that information, because I think 
we should. We have to try to under- 
stand as many of the components as 
we can. but then I think we have to 
go back and put it in the clinical con- 
text. If the patients are doing all right 
and our numbers say they aren't, 
then we could do more harm by try- 
ing to manipulate the situation. So. I 
think we need to try to understand it. 
We need better methods of evalua- 
tion and monitoring, but we have to 
put it all into a clinical context. 

Wood: Lactic acidosis is what I use 
to indicate inadequate oxygenation. 
If my patient has a pH of 7.25 and a 
Pco; of 30 torr, the patient doesn't 
have adequate tissue oxygenation. 
I'm not saying that's the only indi- 
cator, but I don't want us scrambling 
for fantasy here as if we have no in- 
dicators. What I use every day is the 
patient's metabolic acidosis, and as 
soon as I see it I check for a lactate 
level, to make sure I'm not missing 
something else. If it's lactic acidosis, 

I must do something — either increase 
the patient's Do: or decrease the Vo;. 
I do that in every patient except the 
septic patient, because in the septic 
patient the lactic acidosis may not in- 
dicate anaerobic metabolism. Now I 
have a problem that Siegel and Cerra 
gave me 13 years ago.' They showed 
me that in the septic patient, in con- 
trast to the patient with cardiogenic 
or hypovolemic shock, lactic ac- 
idosis isn't an indicator of anaerobic 
metabolism. To solve that problem in 
septic shock is difficult. I need to 
measure the lactate-pyruvate ratio or 
some other indicator of redox po- 
tential, or get an MRI of the organ to 
evaluate whether it's got the right 
level of phosphate or something else. 
These tools are not available yet; so I 
ensure adequate circulating volume 
(Ppcw > 10 torr) and start dobutamine 
(5-10 /jgmin -1 kg -1 ) to ensure ad- 
equate Do: in patients with septic 
shock and lactic acidosis, and I as- 
sume that any persistent lactic ac- 
idosis after this resuscitation will not 
respond to further increases in Do:- 

1. Siegel JH, Cerra FB, Coleman B. 
Giovannini 1, Shetye M, Border JR. et 
al. Physiological and metabolic cor- 
relations in human sepsis (invited 
commentary). Surgery 1979: 186: 163- 

Durbin: I want to comment on the 
last remark and to relate it to oxygen 
content and indices of oxygenation 
as well. An elevated lactate may be a 
problem, but in spite of normal lac- 
tate the body may be experiencing 
anaerobic metabolism in some organ 
beds — anaerobic metabolism that's 
deleterious. So, as with a normal 
P a o:. a normal lactate may give us a 
false sense of security, rather than 
telling us we need to be more ag- 
gressive in treating the problem. 

Cerra: I'd like to echo that point 
again, and even though I think Lar- 
ry's (Wood) precisely stated the 
problem in the inflammatory versus 
the noninflammatory patient, even in 

those patients with inflammatory 
states, a normal lactate does not as- 
sure that all beds are adequately oxy- 
genated. This gets us back to this 
concept of perfusion/diffusion and 
flow distribution, and how do we ac- 
tually measure that in individual or- 

Wood: So. somebody earlier men- 
tioned controversy, and here are the 
areas of controversy. I just don't 
want it to sound as if we don't know 
what to do when we look for an- 
aerobic metabolism. We do know 
what to do. and we do it every day. 
The problem is that when you have 
lactic acidosis, it might not be an- 
aerobic, and when you 'don't got it,' 
it might be anaerobic. In fact, that is 
a problem with a single measurement 
like lactic acidosis. If we could sam- 
ple only the organ that we're inter- 
ested in, then we might have jugular 
venous bulb sampling for cerebral 
hypoxia, and coronary sinus for car- 
diac hypoxia, and renal vein for renal 
hypoxia — another way to go in the 
direction of detecting anaerobic me- 
tabolism, which I think is what we're 
talking about here. Dr Higgins's 
paper was designed to address what 
levels of desaturation we ought to be 
concerned about, as one arm of the 
Do: to avoid the anaerobic metabo- 

Cerra: I think what we really want 
to look at from that end point is 
what's happening with the redox po- 
tential in at least two compartments. 
One is the cytosol, which the lactate- 
pyruvate ratio is looking at. and the 
other is in the mitochondria, and that 
has to do with the proton motor 
force. Currently the only test that's 
clinically useful is the ketone body 
ratio — betahydroxibuturate to aceto- 
acetate — and it's a test that is cum- 
bersome to perform on a real-time 
basis. We also really don't even have 
a real-time assay for pyruvates. A 
number of evaluations of new tech- 
nology to look at these parameters 
are currently in various stages of de- 
velopment and testing. 



When Does Vo 2 Depend on Do 2 ? 

P Terry Phang MD and James A Russell MD 

I. Introduction 

II. Relationship of Yb 2 to Do 2 : Normal vs Pathologic 

III. Clinical Studies Supporting Pathologic Dependence of Vo 2 on Do 2 

IV. Mathematical Coupling in Vo 2 -Do 2 Calculations 

V. What Is the Critical Do 2 in Human Beings and What Is Its 

Relationship to Outcome? 
VI. Summary: When Does Vo 2 Depend on Do 2 ? 


The relationship of oxygen consumption ( Vo 2 ) to 
oxygen delivery (Do 2 ) is central to understanding 
normal cardiopulmonary physiology and the patho- 
physiology of shock and to assessment of tissue is- 
chemia. Knowledge of this relationship is vital for 
constructing a management plan to correctly resus- 
citate patients who have an inadequate Do 2 and 
Vo 2 . In this paper, we review what is known of the 
relationship of Vo 2 to Do 2 from studies in animals 
and humans. Then, we point out our perception of 
methodologic problems that make interpreting data 
from these studies a difficult task. Finally, we end 
our review by pointing out the need for improved 

Dr Phang and Dr Russell are associated with the Program of 
Critical Care Medicine. Pulmonary Research Laboratory, De- 
partments of Surgery and Medicine, St Paul's Hospital and 
University of British Columbia, Vancouver, Canada. 

The preparation of this paper was funded in part by a scholar- 
ship award to Dr Phang from St Paul's Hospital Foundation. 

A version of this paper was presented by Dr Phang on October 
8, 1992. during the Respiratory Care Journal Conference on 
Oxygenation in the Critically 111 Patient held in Puerto Val- 
larta, Mexico. 

Reprints: Dr P Terry Phang. St Paul's Hospital, 1081 Burrard 
St, Vancouver, Canada V6Z 1Y6. 

methods for assessing the presence of tissue is- 
chemia and the adequacy of resuscitation in an in- 
dividual patient. 

Relationship of Vo 2 to Do 2 : 
Normal vs Pathologic 

The relationship of Vo 2 to Do 2 is stylized in Fig- 
ure 1. Point A represents a normal, resting meta- 
bolic state, and Point B represents a state of in- 
creased metabolic rate associated with fever or 
increased activity such as exercise. Increase in Vo 2 
moving from the normal resting state to a state of 
increased metabolic activity is achieved by increase 
in Dr> — usually from an increase in blood flow and 



P« Critical Do 2 

Do 2 

Fig. 1 . The normal relationship of Vo 2 to D02. Point A rep- 
resents a normal resting metabolic state; Point B repre- 
sents a state of increased metabolic rate. At Point C, 
where D02 falls below the critical D02. V02 is dependent 
on Do2- 




to a lesser extent from increased tissue : extrac- 
tion. During states of increased metabolic activity, 
the relationship between Vo 2 and D02 really is that 
of Do: dependence on Vo 2 : Blood flow and tissue 
2 extraction increase in response to an increase in 
metabolic demand. Conversely, if Do: is reduced 
from normal by a decrease in cardiac output (CO.) 
or a decrease in arterial oxygen content, as in states 
of shock or hypoxia, Vo 2 supporting the basal aero- 
bic metabolic state is defended by increasing tissue 
2 extraction. At Point C, where Do 2 falls below 
the critical Do 2 , Vo 2 falls below the level associated 
with the basal metabolic state because increase in 
2 extraction by the tissue is not adequate to com- 
pensate for the decrease in Do 2 - Here, Vo 2 is de- 
pendent on Do 2 , and with Vo 2 below the level as- 
sociated with basal aerobic metabolism, tissue is 

Experimental data that support this normal re- 
lationship of Vo 2 to Do 2 are derived from studies by 
Cain 1,2 of whole-body Vo 2 and Do 2 in anesthetized 
dogs. Cain identified a stable plateau phase (like 
Point A in Figure 1) and a steeply sloped phase (like 
Point C in Figure 1 ) of the relationship between Do 2 
and Vo 2 and determined the critical Do 2 indicating 
onset of dependence of Vo 2 on Do 2 and onset of tis- 
sue ischemia as 9.8 mL 2 • min "' kg 1 . It is im- 
portant to note that the critical Do 2 was the same for 
decreases in Do 2 achieved by hypoxic hypoxia (de- 
creasing Po 2 ). by anemic hypoxia (hemodilution). 
or by stagnant hypoxia (decreased blood flow). Al- 
though the 2 extraction ratio (Vo 2 /Do 2 ) for any 
given Do 2 is greater for stagnant hypoxia than for 
hypoxic hypoxia, the onset of ischemia, or the crit- 
ical Do 2 . is dependent on the total amount of 2 ex- 
tracted by the tissues and independent of type of 
hypoxia. 3 Note that 2 extraction increases as Do 2 
decreases in all types of hypoxia. However, in stag- 
nant hypoxia the arterial-venous 2 content differ- 
ence increases as cardiac output decreases, whereas 
in hypoxic or anemic hypoxia the arterial-venous 
2 content difference does not increase, and both 
arterial and venous 2 content decrease. 

Subsequent to this initial work by Cain, we and 
others have confirmed the biphasic nature of the re- 
lationship of Vb 2 to Do 2 . 4 "'" In addition, the notion 
that the critical Do 2 heralds the onset of tissue is- 
chemia is supported by evidence of onset of rise of 
systemic lactate that slightly precedes the critical 

Do 2 . 4 ' 6 * 7 As well, Gutierrez et al s showed that a de- 
crease in Vo 2 of muscle was associated with an in- 
crease in lactate and inorganic phosphate and a de- 
crease in phosphocreatine indicating that 2 supply 
limitation is associated with anaerobic ATP (aden- 
osine triphosphate) production. 

Effect of temperature on the Vo 2 -Do 2 relation- 
ship is described by Schumacker et al 9 and by Gu- 
tierrez et al.'" Both investigators confirmed that hy- 
pothermia decreases Vo 2 associated with the basal 
metabolic state, and both reported that critical Do 2 
is lower for hypothermia of 30 to 34°C at 5-6 mL 
2 • min "' • kg "' compared to normothermia at 7-9 
mL 2 • min -1 ■ kg ' (that is, hypothermia delays the 
onset of ischemia). However, the effect of hypo- 
thermia on the 2 extraction ratio at the critical Do 2 
(Vo 2 /Do 2 or arterial- venous 2 content difference/ 
arterial 2 content) is unclear from these studies 
because Schumacker et al reported that hypother- 
mia is associated with lower 2 extraction ratio at 
the critical Do 2 . but Gutierrez et al found no differ- 
ence in 2 extraction ratio at the critical Do 2 be- 
tween hypothermic and normothermic animals. 

Effect of P 5U (ie, 2 tension at 50% S a o:) on the 
critical Do 2 was investigated by Schumacker et al" 
who found that a lower P 5n from transfusion with 
low P 50 erythrocytes was not associated with higher 
critical Do 2 . Although P 50 did not affect onset of is- 
chemia, low P 50 did affect maximal 2 extraction at 
very low Do 2 . These results support previous find- 
ings by Cain and Adams' 2 in which infusion of 
NaHCO, (inferred low P 50 ) decreased 2 extraction 
at levels of Do 2 below the critical Do 2 - 

Sepsis increases the critical Do 2 and decreases 
the 2 extraction ratio at the critical Do 2 in animal 
models of sepsis, including pseudomonas bac- 
teremia and E coli endotoxin. 613 The critical Do 2 
was increased by sepsis (11-13 mL 2 ■ min" ' • kg -1 ) 
compared to controls that were not septic (7-8 
niLOi • min"' • kg"'). As well, septic animals had 
lower 2 extraction at the critical Do 2 (0.51-0.54) 
than did nonseptic controls (0.71-0.78). Baseline 
Vo 2 associated with the basal metabolic state was ele- 
vated in septic animals (7 mL 2 ■ min" 1 ■ kg"') com- 
pared to nonseptic controls (6 mL 2 ■ min" 1 ■ kg"' ). 

A stylized comparison of nonseptic and septic 
relationships of Vo 2 to Do 2 is given in Figure 2. In 
these animal studies, the term "pathologic de- 
pendence of Vo 2 on Do 2 " refers to finding that sep- 





, r * 

/ y 


1 Normal 





Do 2 

Fig. 2. Comparison of the pathologic ( ) and normal 

( ) relationships of Vo 2 to Do 2 . "Pathologic depen- 
dence of V02 on D02" refers to the finding that critical D02 
is higher and O2 extraction at the critical D02 is lower in 

sis is associated with earlier onset of ischemia in- 
dicated by higher critical Do 2 and lower : extrac- 
tion at the critical Do 2 . Impaired tissue 2 extrac- 
tion during sepsis was demonstrated in gut by Nel- 
son et al 13 and our group 14 and in skeletal muscle 
by Bredle et al. 15 Impaired microcirculatory regu- 
lation of Do:-matching to Vo 2 as demonstrated by 
impaired vascular reactivity 13 and by widened dis- 
tribution of red blood cell transit times 14 in gut 
could explain the impairment of tissue 2 extrac- 
tion associated with sepsis. Impaired microcircu- 
latory regulation of Do 2 -matching to Vo 2 may be 
due primarily to endothelial response to sepsis and 
to a lesser extent from microemboli of platelets and 
leukocytes within the capillary bed. 1617 

The critical Do 2 is not altered by acute lung 
injury. Pepe and Culver 18 found no significant 
difference in the critical Do 2 between dogs given 
oleic acid, a model of acute lung injury (14 
mL 2 • min -1 • kg -1 ), and control animals (12 
mL0 2 min ' kg' 1 ). Long et al 19 foud that oleic- 
acid-injured dogs had higher critical Do 2 than con- 
trols (9.5 vs 7.4 mLO: • min -1 ■ kg'), but coexistent 
hyperoxia in this study confounds interpretation of 
the independent effect of oleic acid on the critical 
Do 2 . From these studies, it seems that sepsis but 
not acute lung injury without sepsis increases the 
critical Do 2 - 

Clinical Studies Supporting Pathologic 
Dependence of Vo 2 on Do 2 

Compared to data from animal studies, data from 
human studies supporting normal and pathologic 
relationships of Vo 2 to Do 2 are much less clear. 

Whereas normal and pathologic relationships of 
Vo 2 to Do 2 are demonstrated by finding differences 
in the critical Do 2 in animal studies, in human stud- 
ies a normal relationship of Vo 2 to Do 2 is demon- 
strated by finding no change in Vo 2 in response to 
manipulation of Do 2 , and a pathologic relationship 
of Vo 2 to Do 2 has been interpreted as finding a sig- 
nificant change in Vo 2 in response to manipulation 
to change Do 2 - As such, the definition of patholog- 
ic dependence of Vo 2 on Do 2 used in clinical stud- 
ies differs substantially from the definition used in 
animal studies because the critical Do 2 and 2 ex- 
traction ratio at the critical Do 2 are not measured in 
human studies. Many clinical studies of patients 
with adult respiratory distress syndrome (ARDS) 
and sepsis have shown that Vo 2 changes in response 
to changes in Do 2 produced by PEEP, 20 " 23 fluid (in- 
cluding crystalloid, colloid, and blood), 24 ' 29 and 
vasoactive medications.- 10 ' 31 This finding has been 
interpreted to mean that Do 2 is inadequate and that 
tissue ischemia is present. The elevation of plasma 
lactate in association with dependence of Vo 2 on 
Do 2 has substantiated the inference that tissue is is- 
chemic. 34 " 37 

Mathematical Coupling in 
Vo 2 -Do 2 Calculations 

Concern arises from clinical studies that report 
finding pathologic dependence of Vrj 2 on Do 2 - Not 
only has a pathologic critical Do 2 not been defined 
despite the achievement of very high levels of 
Do 2 , 38 " 40 but also the methodology in these studies 
is of concern. The methodology used to demon- 
strate pathologic dependence of Vo 2 on Do 2 em- 
ploys calculation of Vo 2 and Do 2 from cardiac out- 
put (CO.) measured using a thermodilution pul- 
monary artery catheter and oxygen content measure- 
ments of arterial (C a o 2 ) and mixed venous (Cvo 2 ) 

Do: = CO.* Cao:. V 0: = CO.* (C a o; - Co:). 

Because measurements of CO. and C a o 2 are shared 
between the independent variable Do : and the de- 
pendent variable Vo 2 , mathematical coupling of 
shared measurement errors can result in false cor- 
relation between Vo 2 and Do 2 . In other words, the 
use of this methodology makes the finding of a de- 
pendence of Vo 2 on Do 2 suspect. 

When mathematical coupling is avoided in stud- 
ies of the relationship of Vo 2 to D02 °y direct meas- 




urement of Vo 2 from respiratory gas analysis, we 
and others have not found pathologic dependence of 
Vo : on Do 2 . 41 47 In these studies, patients had ARDS 
and sepsis, and in general were comparable in sever- 
ity of illness (as indicated by similar APACHE II 
score, Fio 2 , l eve l of PEEP, and elevation of lactate) 
to patients in the other studies that report finding 
pathologic dependence of Vo 2 on Do 2 . As well, the 
method of changing Do 2 (including PEEP, blood 
transfusion, and dobutamine) was similar. 

Does mathematical coupling account for finding 
pathologic dependence of Vo 2 on Do 2 ? Some of the 
studies that directly measure Vo 2 from respiratory 
gases also report simultaneous Vo 2 from calcula- 
tions using CO. and C a o 2 . 41 " 43 ' 46 In these studies, 
Vo 2 calculated using CO., C a o 2 , and Cv0 2 changed 
significantly after a change in Do 2 , whereas direct- 
ly measured Vo 2 from respiratory gases did not 
change. Because the measurements of Vo 2 are in the 
same patient and are recorded simultaneously dur- 
ing each change in Do 2 , some difference in meth- 
odology is responsible for finding difference in 
change of Vo 2 between methods of determining 
Vo 2 . We are confident that our direct measurements 
of Vo 2 from respiratory gases are accurate (with 
overall measurement error of 1.8%) and are sensi- 
tive to detect a 10% change in Vo 2 , as we have 
shown. 4849 In addition, in our studies we minimized 
spontaneous changes in Vo 2 in patients by use of 
sedation and paralyzing medications as clinically 
required, and body temperature did not change dur- 
ing the 2-4 hour period of each patient study. As a 
result, coefficient of variation for direct measure- 
ment of Vo 2 from respiratory gases ranged from 4% 
at Fio 2 < 0.55 to 7% at Fio 2 of 0.8. Because our 
measurement of Vo 2 from respiratory gases is ac- 
curate and sensitive and because spontaneous 
change in Vo 2 is minimized, we are confident that 
no change occurred in directly measured Vo 2 from 
respiratory gas analysis. Moreover, there was sig- 
nificant and marked increase in Do 2 (upward of 
20%) using blood transfusion and dobutamine dur- 
ing which directly measured Vo 2 did not change. 41 " 43 
We next examined possible sources of systematic 
elevation in calculated Vo 2 as cause of possible ar- 
tifactual change in calculated Vo 2 . Artifactual el- 
evation of calculated Vo 2 could result from eleva- 
tion in CO. and C a o 2 and depression in Cv02- none 
of which have been found except for one report in 
which CO. was overestimated at high ranges. 50 
Therefore, we suggest that mathematical coupling 

is the most likely source of systematic elevation of 
calculated Vo 2 . In the future, we plan to investigate 
the effect of mathematical coupling on the relation- 
ship between calculated Vo 2 and Do 2 using meth- 
ods proposed by Stratton et al 51 that define the con- 
tribution of the slope of random measurement er- 
rors of Vo 2 on Do 2 and the reliability coefficient of 
Do 2 on the observed slope. 

What Is the Critical Do 2 in Human Beings and 
What Is Its Relationship to Outcome? 

The critical Do 2 in normal humans has not been 
reported in the literature. However, three studies 52 " 54 
report the critical Do 2 in hypothermic patients un- 
dergoing cardiopulmonary bypass. In these studies, 
Vo 2 supporting the basal aerobic metabolic state 
was lower than normal, consistent with hypo- 
thermia. Because hypothermia results in lower ba- 
sal Vo 2 and may interfere with 2 extraction, 9 the 
critical Do 2 reported in these studies (about 8-9 
mLO; ■ min" 1 • kg" 1 ) may not reflect the normal 
critical Do 2 . The 2 extraction ratio at the critical 
Do 2 in these studies was approximately 0.35, which 
is much lower than the 2 extraction ratio at the crit- 
ical Do 2 of normal animals (approximately 0.7). As 
well, the critical Do 2 in two of the studies 5354 was 
determined by regression of pooled patient data 
and not individual patient data, which could be mis- 
leading. The critical Do 2 was reported in two other 
studies as 21 mL0 2 ■ min"' ■ kg"' in ARDS pa- 
tients 22 and as 15 mL0 2 min"' -kg" 1 in septic pa- 
tients. 55 However, the critical Do 2 derived in these 
studies could be misleading because fewer data 
points are present above the critical Do 2 and be- 
cause the method of determining the critical Do 2 in 
these studies was by regression of pooled patient 
data and not individual patient data. Pooling of pa- 
tient data ignores potentially important differences 
between patients, such as temperature, body size in 
relationship to active metabolic tissue and distri- 
bution of blood flow to various organs, and effect 
of illness and injury. 

Relationship of outcome to the critical Do 2 is un- 
clear in part because the critical Do 2 in individual 
patients is difficult to determine. Because many 
factors that influence the critical Do 2 may vary (in- 
cluding varying basal Vo 2 and altered ability of tis- 
sues to extract 2 due to illness or injury), a range 
of critical Do 2 in individual patients is likely. De- 
spite this uncertainty of exact value of the critical 




D02 in individual patients, many studies show that 
survivors of critical illnesses have higher values for 
Do: and Vo 2 than nonsurvivors and that non- 
survivors have associated higher values of lac- 
tate. 55 " 60 Relationship of outcome to the finding of 
dependence of calculated Vo 2 on Do 2 is also un- 
clear. Some have found dependence of \fo 2 on Do 2 
to be associated with poor outcome and absence of 
dependence of Vo 2 on Do 2 to be associated with 
good outcome, 61 - 62 whereas we found significant 
change in calculated Vo 2 in response to Do 2 in both 
survivors and nonsurvivors when we retro- 
spectively examined our ARDS patients. 63 In fact, 
we found that for nearly equivalent change in Do 2 
of about 25%, survivors had an increase in calculat- 
ed Vo 2 of about 18%; whereas nonsurvivors had an 
increase in calculated Vo 2 of 8%, which is opposite 
to findings in other studies. 

It is worthy to note that those studies that relate 
Do 2 and Vo 2 to outcome report group statistics and 
that two questions arise in application of findings 
from group statistics to the individual patient. (1) 
What does a group average value for Do 2 of 600 
mL 2 ■ min" 1 ■ m~ 2 (about 15 mL 2 • min" 1 ■ kg" 1 ) 
for survivors as suggested by Shoemaker et al 64 
mean relative to the suggestion of a target resus- 
citation level of Do 2 in the individual patient? (2) 
Does finding an increase in Vo 2 in response to an 
increase in Do 2 in an individual patient mean that 
the patient has ischemic tissue and that more Do 2 is 
required? To answer the first question, in Figure 3 
we present retrospective data from our studies of 34 
nonsurvivors and 21 survivors of ARDS in whom 
nonvasodilating interventions (PEEP, blood trans- 
fusion, dobutamine, elevation of legs) were used. 
Calculated Vo 2 is plotted on the Y-axis in response 
to change in Do 2 plotted on the X-axis. Note that 
average values of Vo 2 and Do 2 are significantly 
greater for survivors than for nonsur-vivors. How- 
ever, on an individual basis, because of a wide 
spread of the data above and below the average val- 
ue, nearly as many survivors have Do 2 < 600 mL 
2 min~'m 2 as have > 600 mL 2 ' min" 1 ■ nr 2 . 
Similarly, nearly as many nonsurvivors have 
Do 2 < 600 mL 2 ■ min ' m 2 as have > 600 mL 
2 min ' -m 2 since there is no clear separation of 
nonsurvivors and survivors who have Do 2 below 
and above 600 mL 2 ■ min ' ■ nr 2 , respectively. 
The results do not allow predicting survival using 

the group mean for Do 2 of 600 mL 2 ■ min" 1 • m" 2 . 
Furthermore, this value of Do 2 does not indicate 
presence or absence of tissue ischemia or adequacy 
of resuscitation of an individual patient. 



f 300 

200 ■ 

> 100 

600 1200 

Do 2 (mL • min- 1 • nr 2 ) 






D02 (mL* min- 


Fig. 3. Relationship of calculated Vo 2 to D02 in survivors 
and nonsurvivors of ARDS. Our analysis does not reveal 
a difference in response of calculated Vo 2 to D02 based 
on survival. 

In answer to the second question — Does finding 
an increase in Vo 2 in response to an increase in Do 2 
in an individual patient mean that the patient has is- 
chemic tissue and that more Do 2 is required? — no 
established statistical method can determine wheth- 
er changes in two measurements of Do 2 and Vo 2 in 
an individual patient are significantly different. 
However, finding dependence of Vo 2 on Do 2 could 
mean that an important change in V02 is observed 




in response to an important change in D02 in an in- 
dividual patient. We suggest that an important 
change is a change in a variable more than twice its 
coefficient of variation because this change is sim- 
ilar to a 96% confidence interval at 2 standard de- 
viations from the mean of a normal sample. For ex- 
ample, if the coefficient of variation of D02 calcu- 
lated from CO. and C a o 2 is 5-10%, then a change 
in D02 > 10-20% is suggested to be important. 
Coefficient of variation for D02 is estimated from 
coefficients of variation for CO. of 5-10%, hemo- 
globin concentration 1%, arterial oxygen saturation 
0.3%, and arterial oxygen tension 5%. 65 Similarly 
if the coefficient of variation for calculated Vo 2 is 
estimated as 8-12%, slightly greater than that for 
D02 because of variation in measurement of Cv02 
based on coefficient of variation for mixed venous 
oxygen saturation of 1 % and mixed venous oxygen 
tension 4%, then a change in Vo 2 > 16-24% might 
be considered important. However, even if this pro- 
posed technique to evaluate importance of change 
in D02 and calculated Vo 2 is acceptable, the prob- 
lem of mathematical coupling remains. It is worthy 
to note that the slope of shared measurement errors 
has been estimated by Stratton et al 51 as 0.24 and 
0.29 based on studies by Powers et al 2n and Danek 
et al. 21 These values for the slope of shared meas- 
urement errors were greater than the observed 
slope of the relationship of Vo 2 on D02 such that 
mathematical coupling of shared measurement er- 
rors influences the observed slope in a positive di- 
rection. That is, mathematical coupling spuriously 
increases the observed slope in the V02-D02 re- 
lationship. In addition, these observations of Vo 2 
and D02 assume that no spontaneous changes in 
V02 and D02 (occurring as a result of patient illness 
or activity) confound interpretation of change in Vo : 
as a marker of tissue ischemia in an individual pa- 
tient. 6667 Perhaps further contradicting use of change 
in calculated Vo 2 in response to change in D02 as a 
marker of tissue ischemia in an individual patient is 
the observation that a plateau phase in calculated 
V02 has not been reported in individual patients 
when D02 is increased to very high levels. 38 " 40 

Based on the preceding discussion, we suggest 
that setting a goal for D02 resuscitation of 600 mL 
2 min"' -m" 2 and observing significant change in 
calculated Vo 2 in response to change in D02 have 
limited utility as markers of tissue ischemia and ad- 

equacy of resuscitation in individual patients. We 
continue to assess whole-body ischemia and ade- 
quacy of resuscitation by assessing mentation, 
urine output, skin perfusion, and lactate but rec- 
ognize that these variables are not specific for is- 
chemia. Perhaps 2 extraction ratio, gastric intra- 
mucosal pH, or other measurements will prove to 
be sensitive and specific markers of tissue ischemia 
and adequacy of resuscitation. 68 " 7 ' 

Summary: When Does Vo 2 Depend on D02? 

The relationship of Vo ; to D02 is well character- 
ized in animal studies as a biphasic relationship, with 
an independent phase indicating aerobic metabo- 
lism and a dependent phase indicating anaerobic 
metabolism and tissue ischemia. From such animal 
studies, the critical D02 indicating onset of tissue 
ischemia is about 7-9 mL 2 ■ min"' ■ kg"'. Oxygen 
extraction ratio at the critical D02 is about 0.7 and 
corresponds to the slope of the dependent phase of 
the V02-D02 relationship. In animal models of sep- 
sis, the critical D02 is increased to about 11-13 
mL 2 ' min "' ■ kg"' and the 2 extraction ratio at 
the critical D02 decreases to about 0.5. 

In contrast to animal studies, the relationship of 
V02 to D02 is not well characterized in human stud- 
ies. Information on the critical Do 2 is preliminary 
such that values for the critical Do 2 and 2 extrac- 
tion ratio at the critical D02 or slope of the depen- 
dent phase of the V02-D02 relationship are uncer- 
tain at this time. Pathologic dependence of Vo 2 on 
Do 2 has been reported in septic ARDS patients, 
with slopes ranging from 0.1 to 0.3. In such stud- 
ies, methodologic error is possible from mathemat- 
ical coupling of shared measurement errors be- 
tween calculated Vo : and Do 2 — especially because 
no dependence has been reported when directly 
measured Vo 2 from respiratory gas analysis has 
been used in similar studies. Methodologic error in 
studies that have demonstrated pathologic depen- 
dence of Vo 2 on Do 2 is further suspect because an 
independent phase of Vo 2 on D02 has not been ob- 
served in individual patients at high levels of D<>- 
Alternative explanations suggested by Dantzker et 
al 72 for finding a sloped relationship of Vo 2 to Do 2 
are that the initial resting Do 2 is dangerously close 
to the critical Do 2 or that spontaneous changes in 
Vq 2 and Do 2 occur during measurements. As well, a 




small slope (estimated at less than 10%) may occur 
above the critical Do 2 due to organs that have 
known positive relationships between blood flow 
and 2 demand, such as heart, liver, and kidney. 

Lastly, at this time we have no explanation for 
finding differences in response of calculated Vb 2 to 
Do: in two groups of patients (ie, one group dem- 
onstrates dependence of calculated Vb 2 on Do 2 , 
whereas the second group does not). In reviewing 
the data presented in Figure 3, we could not find 
two groups of patients who demonstrate a different 
response of calculated Vo 2 to D02 based on survival, 
presence of sepsis syndrome, or elevation of lactate. 
We speculate that possible reasons for finding dif- 
ferences in response of calculated Vo 2 to D02 in 
studies by others may be due to unrecognized differ- 
ences in patients, severity of illness or injury and 
proximity of initial Do 2 to the critical Do 2 in in- 
dividual patients, and methodologic differences, in- 
cluding differences in magnitude of changes of Do 2 
induced by the intervention, coefficients of variation 
for individual component measurements of calculat- 
ed Do? and V~o 2 , and control of spontaneous chang- 
es in Vo 2 and Do 2 due to change in temperature or 
activity during study. Therefore, we believe that bet- 
ter assessments of the presence of tissue ischemia 
and of the adequacy of resuscitation of the whole 
body and individual organs are required. 


Vital contributions by other members of our team helped 
make this work possible: ICU physicians (Drs Keith R Walley. 
John C Fenwick. Peter M Dodek. Martin G Tweeddale). re- 
search fellows (Drs Juan J Ronco. Kenneth F Cunningham. 
Michael Humer, Byron Friesen. Jean Segal), statistician (Dr 
Barry R Wiggs), nurses, respiratory therapists, and laboratory 
and animal care workers (Mses Diane Minschell and Lynne 


1. Cain SM. Appearance of excess lactate in anesthetized 
dogs during anemic and hypoxic hypoxia. Am J Physiol 
1 965.209(3 ):604-6 10. 

2. Cain SM. Oxygen delivery and uptake in dogs during 
anemic and hypoxic hypoxia. J Appl Physiol 1977:42 

3. Gutierrez G, Pohil RF. Strong R. Effect of tlow on O: 
consumption during progressive hypoxemia. J Appl 
Physiol 1988:65(2):601-607. 

4. Segal J, Phang PT, Walley KR. Low-dose dopamine 
hastens the onset of gut ischemia in a porcine model of 

hemorrhagic shock. J Appl Physiol 1992;73(3):1 159- 

5. Cilley RE. Polley TZ Jr. Zwischenberger JB. Toomasian 
JM. Bartlett RH. Independent measurement of oxygen 
consumption and oxygen delivery. J Surg Res 1989:47 

6. Nelson DP. Beyer C, Samsel RW. Wood LDH. Schu- 
macker PT. Pathologic supply dependence of O: uptake 
during bacteremia in dogs. J Appl Physiol 1987:63(4): 

7. Friesen B, Phang PT. Humer M. Walley KR. Does gut 
become ischemic before the whole body during hemor- 
rhagic shock? (abstract). Am Rev Respir Dis 1992:145 
(4. Part2):A791. 

8. Gutierrez G. Pohil RJ. Narayana P. Skeletal muscle O; 
consumption and energy metabolism during hypoxemia. 
J Appl Physiol 1 989:66(5 ):2 1 1 7-2 1 23. 

9. Schumacker PT. Rowland J. Saltz S. Nelson DP. Wood 
LDH. Effects of hyperthermia and hypothermia on oxy- 
gen extraction by tissues during hypovolemia. J Appl 
Physiol 1987:63(3): 1246-1252. 

10. Gutierrez G Warley AR. Dantzker DR. Oxygen de- 
livery and utilization in hypothermic dogs. J Appl Phys- 
iol 1986:60(31:751-757. 

1 1 . Schumacker PT. Long GR, Wood LDH. Tissue oxygen 
extraction during hypovolemia: role of hemoglobin P50. 
J Appl Physiol 1987:62(5): 1 802- 1807. 

12. Cain SM, Adams RP. O: transport during two forms of 
stagnant hypoxia following acid and base infusions. J 
Appl Physiol 1983;54(6):1518-1524. 

13. Nelson DP, Samsel RW. Wood LDH. Schumacker PT. 
Pathologic supply dependence of systemic and intestinal 
O; uptake during endotoxemia. J Appl Physiol 1988:64 

14. Humer M, Phang PT, Friesen B, Walley KR. Sepsis pre- 
disposes gut to ischemia (abstract). Am Rev Respir Dis 
1992:145(4. Part 2):A315. 

15. Bredle DL. Samsel RW. Schumacker PT. Cain SM. 
Critical O: delivery to skeletal muscle at high and low 
P02 in endotoxemic dogs. J Appl Physiol 1989:66(6): 

16. Samsel RW. Schumacker PT. Oxygen delivery to tis- 
sues. Eur Respir J 1991;4(10): 1258-1267. 

17. Shah DM, Newell JC. Saba TM. Defects in peripheral 
oxygen utilization following trauma and shock. Arch 
Surg 1981:116:1277-1281. 

18. Pepe PE, Culver BH. Independently measured oxygen 
consumption during reduction of oxygen delivery by 
positive end-expiratory pressure. Am Rev Respir Dis 

19. Long GR, Nelson DP, Sznajder JI, Wood LDH, Schu- 
macker PT. Systemic oxygen delivery and consumption 
during acute lung injury in dogs. J Crit Care 1988:3(4): 

20. Powers SR Jr. Manual R. Neclerio M. English M. Man 
C, Leather R. et al. Physiologic consequences of pos- 




itive end-expiratory pressure (PEEP) ventilation. Ann 
Surg l973;178(3):265-272. 

21. Danek SJ. Lynch JP. Weg JG. Dantzker DR. The de- 
pendence of oxygen uptake on oxygen delivery in the 
adult respiratory distress syndrome. Am Rev Respir Dis 

22. Mohsenifar Z. Goldbach P. Tashkin DP. Campisi DJ. 
Relationship between O: delivery and O: consumption 
in the adult respiratory distress syndrome. Chest 1983; 

23. Lorente JA. Renes E, Gomez-Aguinaga MA, Landin L, 
de la Morena J. Liste D. Oxygen delivery-dependent ox- 
ygen consumption in acute respiratory failure. Crit Care 
Med 1991;19(6):770-775. 

24. Rhodes GR. Newell JC. Shah D. Scovill W. Tauber J. 
Dutton RE. et al. Increased oxygen consumption accom- 
panying increased oxygen delivery with hypertonic 
mannitol in adult respiratory distress syndrome. Surgery 

25. Powers SR Jr. Shah D, Ryon D. Newell J. Ralph C. Sco- 
vill W, et al. Hypertonic mannitol in the therapy of the 
acute respiratory distress syndrome. Ann Surg 1977: 

26. Kaufman ESS. Rackow EC, Falk JL. The relationship be- 
tween oxygen delivery and consumption during fluid re- 
suscitation of hypovolemic and septic shock. Chest 

27. Wolf YG. Cotev SM. Perel A. Manny J. Dependence of 
oxygen consumption on cardiac output in sepsis. Crit 
Care Med 1987; 15(3): 198-203. 

28. Lucking SE, Williams TM. Chaten FC. Metz RI. Mick- 
ell JJ. Dependence of oxygen consumption on oxygen 
delivery in children with hyperdynamic septic shock 
and low oxygen extraction. Crit Care Med 1990; 1 8( 12): 

29. Astiz ME. Rackow ER. Falk JL, Kaufman BS. Weil 
MH. Oxygen delivery and consumption in patients with 
hyperdynamic septic shock. Crit Care Med 1987:15(1): 

30. Vincent J-L. Roman A. Kahn RJ. Dobutamine admin- 
istration in septic shock: addition to a standard protocol. 
Crit Care Med 1990:1 8(7 ):689-693. 

31. Bollaert PE. Bauer P. Audibert G, Lambert H, Larcan 
A. Effects of epinephrine on hemodynamics and oxygen 
metabolism in dopamine-resistant septic shock. Chest 

32. Shoemaker WC, Appel PL. Kram HB. Duarte D. Harri- 
er D. Ocampo HA. Comparison of hemodynamic and 
oxygen transport effects of dopamine and dobutamine in 
critically ill surgical patients. Chest 1989:961 1 ): 120- 

33. Mackenzie SJ. Kapadia F, Nimmo GR. Armstrong IR. 
Grant IS. Adrenaline in treatment of septic shock: ef- 
fects on haemodynamics and oxygen transport. In- 
tensive Care Med 1991:17:36-39. 

34. Vincent J-L, Roman A, De Backer D, Kahn RJ. Oxygen 
uptake/supply dependency: effects of short-term do- 
butamine infusion. Am Rev Respir Dis 1990:142:2-7. 

35. Fenwick JC. Dodek PM. Ronco JJ, Phang PT. Wiggs B, 
Russell JA. Increased concentrations of plasma lactate 
predict pathologic dependence of oxygen consumption 
on oxygen delivery in patients with adult respiratory dis- 
tress syndrome. J Crit Care 1990:5(2): 1-6. 

36. Haupt MT, Gilbert EM, Carlson RW. Fluid loading in- 
creases oxygen consumption in septic patients with lac- 
tic acidosis. Am Rev Respir Dis 1 985: 1 31:912-91 6. 

37. Gilbert EM, Haupt MT. Mandanas RY. Huaringa AJ. 
Carlson RW. The effect of fluid loading, blood trans- 
fusion and catecholamine infusion on oxygen delivery 
and consumption in patients with sepsis. Am Rev Respir 
Dis 1986:134:873-878. 

38. Nishimura N. Oxygen conformers in critically ill pa- 
tients. Resuscitation 1984:12:53-58. 

39. Astiz ME, Rackow EC. Kaufman B. Falk JL, Weil MH. 
Relationship of oxygen delivery and mixed venous oxy- 
genation to lactic acidosis in patients with sepsis and 
acute myocardial infarction. Crit Care Med 1988:16(7): 

40. Clarke C. Edwards JD. Nightingale P. Mortimer AJ. 
Morris J. Persistence of supply dependency of oxygen 
uptake at high levels of delivery in adult respiratory dis- 
tress syndrome. Crit Care Med 1991;19(4):497-502. 

41. Ronco JJ. Phang PT. Walley KR. Wiggs B. Fenwick JC. 
Russell JA. Oxygen consumption is independent of 
changes in oxygen delivery in severe adult respiratory 
distress syndrome. Am Rev Respir Dis 1991:143:1267- 

42. Ronco JJ. Fenwick JC, Wiggs BR. Phang PT. Russell 
JA. Tweeddale MG. Oxygen consumption is inde- 
pendent of increases in oxygen delivery by dobutamine 
in septic patients who have normal or increased plasma 
lactate. Am Rev Respir Dis 1993:147:25-31. 

43. Cunningham KF. Phang PT. Ronco JJ Fenwick JC. 
Wiggs BR. Russell JA. Independently measured oxygen 
consumption does not depend on oxygen delivery in 
adult respiratory distress syndrome (abstract). Am Rev 
Respir Dis 1991:143(4. Part 2):A205. 

44. Annat G. Viale J-P. Percival C. Froment M. Motin J. 
Oxygen delivery and uptake in the adult respiratory dis- 
tress syndrome: lack of relationship when measured in- 
dependently in patients with normal blood lactate con- 
centrations. Am Rev Respir Dis 1986:133:999-1001. 

45. Carlile PV. Gray BA. Effect of opposite changes in car- 
diac output and arterial Po: on the relationship between 
mixed venous Po; and oxygen transport. Am Rev Re- 
spir Dis 1989;140:891-898. 

46. Vermeij CG, Feenstra BWA. Braining HA. Oxygen de- 
livery and oxygen uptake in postoperative and septic pa- 
tients. Chest 1990:98(2):415-420. 

47. Lutch JS. Murray JF. Continuous positive-pressure ven- 
tilation: effects on systemic oxygen transport and tissue 
oxygenation. Ann Intern Med 1972:76(2): 193-202. 

48. Phang PT. Rich T, Ronco JJ. A validation and compari- 
son study of two metabolic monitors. JPEN 1990:14(3): 




49. Ronco JJ. Phang PT. Validation of an indirect calori- 
meter to measure oxygen consumption in the ranges 
seen in critically ill patients. J Crit Care 199 1 :6( 1 ):36- 

50. Robbins LD. Dawson JA, Steingrub J, Teres D. Dis- 
crepancy between measured and calculated oxygen con- 
sumption in critically ill patients (abstract). Anesthes- 
iology 1988;69:A169. 

51. Stratton HH, Feustel PJ, Newell JC. Regression of cal- 
culated variables in the presence of shared measurement 
error. J Appl Physiol 1987;62:2083-2093. 

52. Fox LS, Blackstone EH. Kirklin AW, Stewart RW, Sa- 
muelson N. Relationship of whole body oxygen con- 
sumption to perfusion flow rates during hypothermic 
cardiopulmonary bypass. J Thorac Cardiovasc Surg 

53. Shibutani K. Komatsu T, Kubal K, Sanchala V. Kumar 
V, Bizzari DV. Critical level of oxygen delivery in an- 
esthetized man. Crit Care Med 1 984; 1 1(8):640-643. 

54. Komatsu T, Shibutani K, Okamoto K, Kumar V, Kubal 
K, Sanchala V, et al. Critical level of oxygen delivery 
after cardiopulmonary bypass. Crit Care Med 1987; 15 
(3): 194-197. 

55. Tuchschmidt J, Fried J, Swinney R, Sharma OP. Early 
hemodynamic correlates of survival in patients with sep- 
tic shock. Crit Care Med 1989;17(8):7 19-723. 

56. Shoemaker WC, Appel PL. Kram HB. Role of oxygen 
debt in the development of organ failure sepsis, and 
death in high-risk surgical patients. Chest 1992;102(1): 

57. Shoemaker WC, Appel PL, Kram HB, Waxman K, Lee 
TS. Prospective trial of supranormal values of survivors 
as therapeutic goals in high risk patients. Chest 1988; 

58. Moore FA, Haenel JB. Mook EE, Whitehill TA. In- 
commensurate oxygen consumption in response to max- 
imal oxygen availability predicts postinjury multiple or- 
gan failure. J Trauma 1992;33( 1 ):58-67. 

59. Rashkin MC. Bosken C, Baughman RP. Oxygen de- 
livery in critically ill patients: relationship to blood lac- 
tate and survival. Chest 1985;87(5):580-584. 

60. Russell JA, Ronco JJ, Lockhat D, Belzberg A. Kiess M, 
Dodek PM. Oxygen delivery and consumption and ven- 
tricular preload are greater in survivors than in non- 

survivors of the adult respiratory distress syndrome. Am 
Rev RespirDis 1990;141:659-665. 

61. Gutierrez G, Pohil RJ. Oxygen consumption is linearly 
related to O: supply in critically ill patients. J Crit Care 
1986:1(1 ):45-53. 

62. Bihari D. Smithies M. Gimsom A. Tinker J. The effect 
of vasodilation with prostacyclin on oxygen delivery 
and uptake in critically ill patients. N Engl J Med 

63. Ronco JJ. Phang PT. Cunningham KF, Fenwick JC. 
Russell JA. No difference in calculated oxygen kinetics 
after non-vasodilating interventions between nonsur- 
vivors and survivors of ARDS (abstract). Am Rev Re- 
spirDis 1991:143(4, Part 2):A205. 

64. Shoemaker WC, Montgomery ES. Kaplan E. Elwyn 
DH. Physiologic patterns in surviving and nonsurviving 
shock patients. Arch Surg 1973; 106:630. 

65. Mood Am, Graybill FA, Boes DC. Introduction to the 
theory of statistics, 3rd ed. Toronto: McGraw-Hill, 

66. Weissman C, Kemper M. The oxygen uptake-oxygen 
delivery relationship during ICU interventions. Chest 

67. Villar J. Slutsky AS, Hew E, Aberman A. Oxygen trans- 
port and oxygen consumption in critically ill patients. 
Chest 1990;98(3):687-692. 

68. Gutierrez G. Palizas F. Doglio G, Wainsztein N. Gal- 
lesio A, Pacin J, et al. Gastric intramucosal pH as a ther- 
apeutic index of tissue oxygenation in critically ill pa- 
tients. Lancet 1992;339(8787):195-I99. 

69. Gys T, Hubens A, Neels H, Lauwers LF. Peeters R. 
Prognostic value of gastric intramural pH in surgical in- 
tensive care patients. Crit Care Med 1988;16(12):1222- 

70. Doglio GR, Pusajo JF, Egurrola MA, Bonfigli GC, Par- 
ra C, Vetere L, et al. Gastric mucosal pH as a prognostic 
index of mortality in critically ill patients. Crit Care 
Med 1991;19(8):1037-1040. 

71. Gutierrez G, Bismar H. Dantzker DR, Silva N. Com- 
parison of gastric intramucosal pH with measures of ox- 
ygen transport and consumption in critically ill patients. 
Crit Care Med 1992;20(4):45 1-457. 

72. Dantzker DR, Boresman B. Gutierrez G Oxygen supply 
and utilization relationships. Am Rev Respir Dis 1991; 

Phang Discussion 

Hess: Thank you. Terry. I think that 
was probably the most lucid de- 
scription of mathematical coupling 
that I've heard, and I thank you for 
that. I ask the group this morning to 
keep the discussion to pathophysi- 
ology. We'll talk about treatment 
strategies tomorrow afternoon after 
Dr Cerra's talk. 

Cerra: Not being bashful, I do have 
some questions. I have enjoyed read- 
ing your work and I firmly believe in 
this concept of mathematical coup- 
ling. But, I do have some questions 
about your models, and, therefore, 
the strengths of the conclusions that 
come from them. In the dobutamine 
model — yes, you're increasing de- 
livery, but you uncouple delivery 
from blood volume, so that you're 

really not sure you changed the dis- 
tribution of blood flow at all. You 
may be just circulating a smaller 
blood volume faster when, in fact, 
tissue isn't getting the volume of dis- 
tribution that it needs. That's also a 
problem in studies that only change 
cardiac output without taking into 
account the existing vascular capaci- 
tance and the volume of distribution 
to capillary beds. That methodology 




of increasing output without increas- 
ing blood volume is also not what 
we usually do clinically. Especially 
with surgically trained people, their 
first response is usually volume. It's 
based on that concept. Second, I 
have no idea what the metabolic 
state of your patients was. They 
were heavily sedated, and, I think, 
paralyzed. Is that right? 

Phang: The majority of them, 14 of 
them were paralyzed. 

Cerra: Both sedation and paralysis 
suppress oxygen demand; so, I don't 
know where you are in your oxygen 
demand, even if you accept the theo- 
retical concept of supply-depen- 
dency. If you're not in the supply- 
dependent portion of the curve be- 
cause demand is below that, then 
you're not going to see a supply- 
dependent relationship. Third. I 
don't know how to deal with this is- 
sue of mathematical coupling or 
flow-dependent oxygen consumption 
unless it is coupled with some meta- 
bolic measurement that reflects the 
energetic state itself. Last, when the 
same method of determining Vo : (in- 
direct Fick) is applied to two groups 
of patients in the same physiologic 
state and one shows supply-de- 
pendency and one doesn't, how can 
mathematical coupling explain that. 
If it were operative, it should appear 
in both groups. 

Phang: 1 have to apologize. I'm 
really not good at answering multi- 
ple questions. I'll start with what I 
think was the first one. That was, Is 
there a difference between the Vo : - 
D02 relationship, dependent on the 
method of increasing oxygen de- 
livery? Is that right? 

Cerra: I think that's essentially the 

Phang: And so, as a surgeon, you 
said, you'd prefer to see volume in- 
crease instead of flow. You know, 
the other guys, the non-surgeons (I 

hate to use these generalities), for ex- 
ample a guy like Dr Jean-Louis Vin- 
cent (he's not a surgeon, he's one of 
those medical guys), he would prefer 
to see oxygen delivery increased by 
increasing flow as in our current 
study. In fact, we started our V02-D02 
studies by first increasing blood vol- 
ume to increase oxygen delivery. 1 In 
that study the same flat V02-D02 re- 
lationship holds for the method of in- 
creasing Do2- In other words, if you 
use blood volume as a method of in- 
creasing D02 or if you use blood flow 
as a method of increasing D02, the 
same flat relationship of measured 
D02 with a metabolic cart on D02 ap- 
plies. That is, consistently there is no 
sloping of the relationship between 
V02 and Do2- This addresses a little 
bit your second question, the prob- 
lem with controlling the Vo : . 

For the purposes of the meth- 
odology, it's very important to con- 
trol the V02 in order to say that any si- 
multaneously observed slope in the 
V02-D02 relationship in the same pa- 
tient at the same time is not due to 
spontaneous change in Vo 2 . It is a 
methodologic problem of concern. 
How that relates to other markers of 
how critically ill our patients were 
and controlling the Vo : really, then, is 
separated completely from the meth- 
odology. You're shaking your head, 
and I'll let you talk in 1 second. Our 
patients, at the time they're studied, 
are at what we perceive to be past the 
clinical stages of resuscitation. At the 
time of these studies, the patients 
were fully resuscitated clinically, so 
that the V02-D02 studies were meant 
to detect hidden or occult ischemia. 
It's not to say that lactate wasn't 
around because in over half of the pa- 
tients, in 10 out of the 17, at the time 
that the studies were done, the lactate 
was high and the patients were sep- 
tic. Sepsis and lactate are the things 
that some people have used to predict 
finding dependence of Vo : on Do:. 
Despite the fact that patients on an 
individual basis may have been sep- 

tic or had elevated concentrations of 
lactate in their blood, those patients 
still did not demonstrate any de- 
pendence of respiratory-gas-mea- 
sured V02 in response to a significant 
change in Do: at that time point. So, 
they were ill, they were septic, they 
had ARDS, they even had high lac- 
tate, and it turns out that the re- 
lationships of the measured Vo 2 in 
those patients were not sloped and 
were flat. 

I. Ronco JJ, Phang PT. Walley KR. 
Wiggs B. Fenwick JC. Russell JA. 
Oxygen consumption is independent 
of changes in oxygen delivery in se- 
vere adult respiratory distress syn- 
drome. Am Rev Respir Dis 1991; 

Cerra: But you don't know if they 
were anaerobic. That's the point I'm 
trying to get at. Unless you show me 
that at the entrance to the study they 
had excess lactate and. therefore, 
would be at the point in their en- 
ergetics where you would expect to 
see supply-dependency. I can't ac- 
cept the data. The opposite side of 
that coin is in precisely the patients 
you're talking about. This is work 
John Siegel and I did for 6 years be- 
tween 1973 and 1979.' - In septic re- 
suscitated patients. I'll accept your 
definition of resuscitation. A high 
plasma lactate does not necessarily 
imply anaerobic metabolism. And, in 
fact, what you've described are the 
characteristics of aerobic glycolysis, 
which is the normally functioning 
Krebs cycle and normally func- 
tioning electron transport, and not an- 
aerobic metabolism. We must couple 
these physiologic measurements with 
metabolic measurements that reflect 
energetics. Even though I accept the 
argument about coupling, I can't ac- 
cept the second argument about the 
absence of flow-dependency from 
the data until you meet the metabolic 
criteria for electron transport. 




1. Siegel JH, Cerra FB, Coleman B, 
Giovannini I, Shetye M, Border JR. 
et al. Physiological and metabolic 
correlations in human sepsis. Surgery 

2. Cerra FB. Hypermetabolism, organ 
failure and metabolic support. Sur- 
gery 1987:101:1-14. 

Phang: I'm saying that our patients 
were not flow-dependent. Yet, the 
Swan-Ganz measurement may have 
indicated that they were. 

Cerra: I accept that. That's not the 
point I'm getting at. I think that's a 
valid argument. It's the other pieces 
associated with that, and, in part, I'm 
worried about the data because you 
haven't demonstrated to me that 
they "re at a point in their energetics 
at which they should be flow-depen- 
dent. So, the fact that you didn't see 
it in your expired gas analysis 
doesn't tell me it wasn't there. 

Phang: If you're saying that if I had 
a patient who was clearly in shock — 
anyone from a patient with pneu- 
monia and septic shock or a patient 
with hemorrhagic shock from trau- 
ma — and I put my metabolic cart on 
him and then resuscitated him, and I 
had the other measurements to show 
he was in shock (whatever you 
chose, for example. P nuclear mag- 
netic resonance spectroscopy to de- 
termine intracellular ATP/ADP and 
pH), I think that I would agree that 
there would be an increase in that 
patient's Vo?. I think I would observe 
a sloped relationship between V02 
and D02, which I would interpret in 
this instance as the physiologic sup- 
ply-dependence of Vo; on Do:. 
which indicates shock. However, our 
patients were fully resuscitated. I be- 
lieve that our patients were not flow- 

Cerra: Not flow-dependent. OK, I 
think we're at the same point. 

Wood: That's a very interesting dis- 
cussion, and I want to make a couple 

of prospective statements that lead 
me to believe that Frank (Cerra) and 
Terry (Phang) are in violent agree- 
ment. It is crucial to my under- 
standing of what's gone on, and it 
goes like this — if patients are in 
shock, they should be resuscitated. 

Phang: I agree. 

Cerra: I agree. 

Wood: There we are. Now, once you 
resuscitate them to a clinically ac- 
ceptable circulating volume, we have 
been on an uncertain journey con- 
cerning obscure goals of further re- 
suscitation for 10 years if you start 
with Danek, 1 or 20 years if you start 
with Powers 2 — wandering around in 
the intellectual wilderness because 
there was described a relationship be- 
tween Vo; and Do: that was given 
some physiologic significance. We 
just heard as well as it can be said 
that Vo; does not depend on D02 in 
resuscitated patients. If in a large 
group of Canadian septic patients on 
ventilators, there is no flow-depen- 
dence of oxygen consumption, that 
result runs counter to what the prior 
data have been. 14 But we've just 
heard the reasons why it runs coun- 
ter: When you plot one measurement 
against itself, shared measurement er- 
ror makes a relationship when there 
isn't one. 5 You confer significance, 
and you are wrong. So, now, we can 
leave that to rest; those resuscitated 
patients in Canadian ICUs or Minne- 
apolis ICUs, once they're resusci- 
tated, have no flow-dependence. Next, 
there's this Shoemaker approach, 11 
which looks back at some retrospec- 
tive data demonstrating that people 
who have supranormal flows sur- 
vived and those who didn't, didn't, 
and makes a therapeutic approach out 
of it. 

Of course an equally plausible in- 
terpretation is that survivors are able 
to respond to excess volume and va- 
soactive therapy by increasing their 
flow. When this approach (max- 

imizing Do:) was tested prospective- 
ly, 7 there was no difference in sur- 
vivorship between the groups ran- 
domized to this treatment or not (p = 
0.14); yet the authors support this ap- 
proach by a retrospective look at 
both groups showing again that sur- 
vivors had high flows — they don't 
say that their intervention did not 
make it happen. 7 Furthermore, there 
was no relationship, on a retrospec- 
tive look or a prospective look at 
those patients, between D02 and Vo 2 , 
just as Terry Phang demonstrated to- 

The last point that I would like to 
make is that some patients who are 
resuscitated by our usual criteria 
share some characteristics of hyper- 
metabolism or inflammatory states 
(the area that Frank Cerra has been 
looking at for a long time now). 8 " 10 
There is a possibility that there are 
some hidden mechanisms that are 
not characterized by flow-depen- 
dence of D02 — that's what I just saw 
Terry (Phang) present — but are char- 
acterized by some unusual forms of 
accelerated aerobic glycolysis that 
have high lactates and pyruvates (so 
a normal aerobic L-P ratio) that can 
only be picked up by either excess 
lactate, MRI phosphate, or other spe- 
cial tools that not many of us have. 
I've been wishing that I could meas- 
ure pyruvate, but I can't. Frank (Cer- 
ra) does. Measuring the ketone body 
ratio sounds like a great idea, but I 
haven't been able to get our people 
to do it, and most places where I visit 
don't do that either. Those criteria of 
excess lactate are a fishing trip in 
this large body of resuscitated, crit- 
ically ill patients, for something else 
that is unique to their limits of oxy- 
genation, and that uniqueness prob- 
ably has got something to do with 
the inflammatory or septic state it- 
self. 91 " 

I conclude from all that evaluation 
that Terry and Frank are in violent 
agreement. And I really believe it's 
helpful for me not to have to do 




something silly — like maximize the 
oxygen transport, when it has no 
survivorship value — unless there's a 
unique septic patient where I can't 
really be sure that the patient is com- 
pletely resuscitated because (although 
his Do: is 1,200 mL/min and his car- 
diac output is 10 L/min) his blood 
pressure is still only 80 systolic. He 
isn't putting out much urine. His 
temperature, lactic acid, and \b 2 are 
high, and I'm not really sure whether 
I have eliminated anaerobic me- 
tabolism. Yet, when we increased 
Do: further in such septic patients, 
Vo: did not increase in any of them," 
just as Terry Phang demonstrated. 
Because people are septic in ICUs, 
we do need better tools to exclude 
anaerobic metabolism and to block 
the accelerated aerobic glycolysis 
that won't respond to maximizing 
Do:. 9 " 11 

1. Danek SJ. Lynch JP, Weg JG, 
Dantzker DR. The dependence of 
oxygen uptake on oxygen delivery in 
the adult respiratory distress syn- 
drome. Am Rev Respir Dis 1980; 

2. Powers SR Jr, Mannal R, Neclerio 
M, English M, Marr C. Leather R. et 
al. Physiologic consequences of pos- 
itive end-expiratory pressure (PEEP) 
ventilation. Ann Surg 1973:178:265- 

3. Gilbert EM, Haupt MT, Mandanas 
RY, Huaringa AJ, Carlson RW. The 
effect of fluid loading, blood trans- 
fusion, and catecholamine infusion 
on oxygen delivery and consumption 
in patients with sepsis. Am Rev Re- 
spir Dis 1986;134:873-878. 

4. Fenwick JC, Dodek PM, Ronco JJ, 
Phang PT, Wiggs B, Russell JA. In- 
creased concentrations of plasma 
lactate predict pathologic supply de- 
pendence of oxygen consumption on 
oxygen delivery in patients with res- 
piratory distress syndrome. J Crit 
Care 1990;5:81-86. 

5. Stratton HH, Feustel PJ, Newell JC. 
Regression of calculated variables in 
the presence of shared measurement 
error. J Appl Physiol 1987:62:2083- 

6. Shoemaker WC. Appel PL, Kram 
HB, Waxman K, Lee TS. Prospective 
trial of supranormal values of sur- 
vivors as therapeutic goals in high- 
risk surgical patients. Chest 1988; 

7. Tuchschmidt J, Fried J, Astiz M, 
Rackow E. Elevation of cardiac out- 
put and oxygen delivery improves 
outcome in septic shock. Chest 

8. Siegel JH, Cerra FB, Coleman B, 
Giovannini I. Shetye M, Border JR, 
et al. Physiological and metabolic- 
correlations in human sepsis. Surgery 

9. Vary TC. Increased pyruvate de- 
hydrogenase kinase activity in re- 
sponse to sepsis. Am J Physiol 1991; 
260(5, Part 1 ):E669-E674. 

10. Curtis SE, Cain SM. Regional and 
systemic oxygen delivery/uptake re- 
lations and lactate flux in hyperdy- 
namic, endotoxin-treated dogs. Am 
Rev Respir Dis 1992:145(2. Part 1): 

1 1 . Manthous CA, Schmidt GA. Hall JB, 
Pohlman A, Schumacker PT, Samel 
R, Wood LDH. Oxygen consumption 
and delivery in septic shock (ab- 
stract). Am Rev Respir Dis 1993; 
147(4. Part 2):A 616. 

Cerra: I have a second question. It's 
a theoretical sort of argument you 
get. Bihari's data have confused me 
ever since they came out. 1 And I 
think you've taken one approach to 
try to unravel what that observation 
really means. My question is, Is the 
difference between what Bihari did 
and what you did that he gave a di- 
rect microcirculatory, pharmacologic 
vasodilator? And I'm wondering, in 
your study, if you really had a con- 
trol for that? 

1. Bihari D. Smithies M, Gimson A, 
Tinker J. The effects of vasodilation 
with prostacyclin on oxygen delivery 
and uptake in critically ill patients. N 
Engl J Med 1987;317:397-403. 

Phang: No. 

Cerra: And whether you think that 
may make a difference? 

Phang: It may. I don't know the an- 
swer to that. If it tells you that the 
gut opens up after a vasodilator is 
given, and that it is important to re- 
suscitate the gut, he may be right. 
His data, though, when I've looked 
al it, 1 as far as trying to understand 
the potential for coupling, show that 
the nonsurvivor group, the one that 
shows these big jumps in \fo 2 . has 
change in Do: much smaller relative 
to the survivor group. The smaller 
change in D02 could mean a higher 
potential for mathematical coupling 
and an artifactual slope in the non- 
survivor group. But he may have 
something. It may be that you have 
to resuscitate your hypotensive pa- 
tients with vasodilators to see if they 
have occult ischemia. It may be true, 
but I don't know whether his data 
are methodologically correct. 

1. Bihari D. Smithies M. Gimson A, 
Tinker J. The effect of vasodilation 
with prostacyclin on oxygen delivery 
and uptake in critically ill patients. N 
Engl J Med 1987:317:397-403. 

Hudson: I have a question about the 
animal model data that you started 
out with. You said that there were 
animal models of sepsis in which 
there appeared to be flow-depen- 
dence, but in the oleic acid model 
there wasn't flow-dependence. And 1 
think you were referring to the Pepe 
and Culver oleic acid data. 1 and they 
used independent methods of meas- 
uring Vo: and Do> I wonder about 
the septic models. Do you think that 
their use of an independent method 
to avoid the coupling explained 
those results — not whether it was an 
isolated lung injury versus a total 
body inflammatory injury? 

1. Pepe PE, Culver BH. Independently 
measured oxygen consumption dur- 
ing reduction of oxygen delivery by 
positive end-expiratory pressure. Am 
Rev Respir Dis 1985:132:788-792. 

Phang: No. I don't think so. The 
change in Do: when you go from a 




normal or hypermetabolic state to 
zero at the time of death is very, 
very wide. If you see a clear in- 
flection point, I'm not sure that it 
matters that your variables are un- 
coupled or coupled. In the animal 
studies they don't use the term 
"flow-dependence" to describe pa- 
thology. They say "a difference in 
critical D02" or "a difference in the 
extraction ratio at that point." I think 
that the studies in bacteremic and in 
endotoxic dogs' : showed nicely that 
there is a difference between normal 
oxygen extraction and normal onset 
of tissue ischemia versus abnormal. 
I'm not as convinced about the ole- 
ic acid studies. In the study by Pepe 
and Culver, 3 the critical D02 was 12- 
14 mL • min" 1 ■ kg ' in controls and 
oleic-acid groups under normoxic 
conditions. In the study by Long 
and Schumacker and their group, 4 
the critical D02 in hyperoxic oleic- 
acid-treated animals was about 10 
mL • min -1 ■ kg ' compared to hy- 
peroxic control animals at 13 mL • 
min" 1 ■ kg" 1 and normoxic control an- 
imals at 7 mL ■ min" 1 ■ kg" 1 . There- 
fore, although the studies conclude 
that oleic acid had no effect on crit- 
ical Do:, final interpretation may be 
confounded by use of hyperoxia in 
one study and normoxia in the other 
study. I can't see that if you induce a 
generalized systemic inflammatory 
response from oleic acid lung injury 
why it should not show a difference 
in critical oxygen delivery if the 
problem at the tissue level is a prob- 
lem with capillary blood flow mal- 
distribution and inability of the tis- 
sue to extract oxygen. I don't think 
this generalized systemic inflam- 
matory state should be different 
from the septic animals because the 
mediators are probably the same. A 
pure oleic-acid injury to the lung 
should not be different from sepsis 
unless the effect of the oleic-acid in- 
jury is confined to the lung so that 

the inflammation does not become a 
systemic problem. 

1. Nelson DP, Beyer C, Samsel RW, 
Wood LDH, Schumacker PT. Path- 
ological supply dependence of O2 up- 
take during bacteremia in dogs. J Appl 
Physiol 1987;63(4): 1487-1492. 

2. Nelson DP, Samsel RW, Wood LDH. 
Schumacker PT. Pathological supply 
dependence of systemic and intestinal 
O2 uptake during endotoxemia. J Appl 
Physiol 1988;64(6):24 10-24 19. 

3. Pepe PE, Culver BH. Independently 
measured oxygen consumption during 
reduction of oxygen delivery by posi- 
tive end-expiratory pressure. Am Rev 
RespirDis 1985;132:788-792. 

4. Long GR, Nelson DP, Sznajder JI, 
Wood LDH, Schumacker PT. System- 
ic oxygen delivery and consumption 
during acute lung injury in dogs. J Crit 
Care 1988;3(4):249-255. 

Wood: I just want to reply to Terry's 
question that our same group of peo- 
ple (Schumacker et al) who showed 
the endotoxin 1 and bacteremic 2 path- 
ologic supply-dependence in dogs 
showed no such phenomenon in dogs 
with oleic-acid lung injury by the 
same uncoupled methods. 1 Those 
data lead that group to think that sep- 
sis is different from acute lung injury, 
though we often lump them, and in 
fact they've been lumped from the 
earlier clinical studies of Powers, 4 
Danek, 5 and others. 6 - 7 Many patients 
with ARDS are septic, and so that's 
where a confusion has arisen in inter- 
preting the clinical data. I think the 
value of actually finding a pathologic 
supply-dependence in animals is that 
it gives additional tools to exclude 
this phenomenon in patients. It lets us 
get at this absence of flow- 
dependence in patients who are sep- 
tic — because when push-comes-to- 
shove, we still don't understand what 
is going on in the effective resus- 
citation from tissue hypoxia in the 
septic patients. The animals clearly 
showed this unusual pathologic sup- 

ply-dependence and extraction de- 
fect, which we tried to attribute to 
our patients but our patients don't 
have them. The septic dogs have an 
inflammation; we did that. And sep- 
tic, acidotic patients have an in- 
flammation. There's something meta- 
bolic here that is not necessarily an- 
aerobic metabolism, and that's what 
Frank's (Cerra) been saying. 

1. Nelson DP, Beyer C. Samsel RW. 
Wood LDH, Schumacker PT. Path- 
ological supply dependence of O2 up- 
take during bacteremia in dogs. J 
Appl Physiol 1987;63:1487-1492. 

2. Nelson DP. Samsel RW. Wood LDH, 
Schumacker PT. Pathological supply 
dependence of systemic and intestinal 
O2 uptake during endotoxemia. J 
Appl Physiol 1988;64:2410-2419. 

3. Long GR, Nelson DP, Sznajder JI. 
Wood LDH, Schumacker PT. Sys- 
temic oxygen delivery and consump- 
tion during acute lung injury in dogs. 
J Crit Care 1988;3:249-255. 

4. Powers SR Jr. Mannal R, Neclerio M. 
English M. Marr C, Leather R. et al. 
Physiologic consequences of positive 
end-expiratory pressure (PEEP) ven- 
tilation. Ann Surg 1973;178:265-272. 

5. Danek SJ, Lynch JP. Weg JG, Dantz- 
ker DR. The dependence of oxygen 
uptake on oxygen delivery in the 
adult respiratory distress syndrome. 
Am Rev Respir Dis 1980:122(3):387- 

6. Gilbert EM. Haupt MT, Mandanas 
RY. Huaringa AJ. Carlson RW. The 
effect of fluid loading, blood trans- 
fusion, and catecholamine infusion 
on oxygen delivery and consumption 
in patients with sepsis. Am Rev Re- 
spir Dis 1986:134:873-878. 

7. Fenwick JC. Dodek PM, Ronco JJ. 
Phang PT. Wiggs B. Russell JA. In- 
creased concentrations of plasma lac- 
tate predict pathologic supply de- 
pendence of oxygen consumption on 
oxygen delivery in patients with res- 
piratory distress syndrome. J Cril 
Care 1990;5:81-86. 



Assessment of Oxygenation: Oxygenation Indices 

Loren D Nelson MD 








Historical Assessment of Oxygenation 
Determinants of Clinical Utility of Indices 
Factors Affecting Alveolar Oxygen Tension 
Identification, Derivation, and Application of Indices 
of Oxygenation 

A. Tension-Based Indices 

B. Saturation- and Content-Based Indices 
Future Direction of Oxygenation Monitoring 
In Conclusion 


A number of indices are used to describe oxy- 
genation in critically ill patients. This paper re- 
views the commonly used oxygenation indices and 
critically assesses their application in patient care. 

Historical Assessment of Oxygenation 

Before the late 1950s, the assessment of oxyge- 
nation was usually made by observation of the pa- 
tient's overall condition and skin color. The di- 
agnosis of hypoxia was most commonly associated 
with the physical finding of cyanosis — a finding 
that generally requires the presence of more than 5 
g/dL of deoxyhemoglobin in arterial blood. 1 The 
finding of cyanosis, therefore, reflects severe arteri- 

Dr Nelson is Associate Professor of Surgery and Anes- 
thesiology, and Director of the Surgical Intensive Care Unit at 
Vanderbilt University. Nashville, Tennessee. 

A version of this paper was presented by Dr Nelson on Oc- 
tober 8, 1992. during the Respiratory Care Journal Confer- 
ence on Oxygenation in the Critically 111 Patient, held in Puerto 
Vallarta, Mexico. 

Reprints: Loren D Nelson MD. Associate Professor of Surgery 
and Anesthesiology. Vanderbilt University, T-2104 Medical 
Center North. Nashville TN 37232-2100. 

al desaturation in the presence of a normal hemo- 
globin (Hb) concentration and is almost un- 
detectable in patients with severe anemia. 

The development of the Clark P02 electrode in 
the 1950s, together with the widespread application 
of this device in the 1960s, has been heralded as 
one of the landmark events leading to modern crit- 
ical care. 2 Prior to the development of the P02 elec- 
trode, quantitative assessment of oxygenation was 
possible only by using the complex and time- 
consuming Van Slyke method of determining the 
total oxygen content of whole blood. 

In the 1960s and 1970s, arterial oxygen tension, 
or partial pressure, (PaO:) became the standard by 
which arterial oxygenation was assessed, although 
clinicians realized that P a o2 alone gave little in- 
formation regarding tissue oxygenation. 

Throughout the 1970s, other indices were de- 
veloped in the hope that they would better reflect 
tissue oxygenation. Blood gas analyzers were fitted 
with software that calculated hemoglobin oxygen 
saturation from a normal oxyhemoglobin dissocia- 
tion curve modified by the patient's actual pH a and 
arterial carbon dioxide tension (PaCO:)- The cal- 
culated arterial oxygen saturation (S a 0:) could then 
be used with the patient's most recent hemoglobin 
concentration to calculate the volumetric content of 
oxygen in the patient's arterial blood (do:)- 1 Spec- 
trophotometry devices (CO-oximeters) provided 




direct determination of S a o> dyshemoglobins. and 
total hemoglobin (tHb). 

Determinants of Clinical Utility of Indices 

The clinical utility of an oxygen variable or an 
index depends on two factors: ( 1 ) the information it 
provides regarding the adequacy of tissue oxygena- 
tion, and (2) the information it provides that can af- 
fect therapeutic interventions. The therapeutic deci- 
sion-making utility of the variable is determined by 
the guidance it provides the clinician in differen- 
tiating among a number of factors that may be clin- 
ically altered — including the gas source, the lungs, 
the blood, and the tissues. 

Factors Affecting Alveolar Oxygen Tension 

Barometric pressure (P B ) and inspired oxygen 
fraction (F102) determine the inspired tension of ox- 
ygen (Pio:)- According to Dalton's law, the P102 is 
reduced by the partial pressures of other gases 
present, including those added by the respiratory 
system. Typically, a small amount of water vapor is 
added by evaporation from the proximal airways, 
and C0 2 is added by diffusion from capillary blood 
into perfused alveoli. Therefore, the alveolar oxy- 
gen tension (PAO2) is less than the P102 and is ap- 
proximated by the equation 



where R = respiratory exchange ratio (CO: excretion/O: 
uptake by the lungs). 

Identification, Derivation, and Application 
of Indices of Oxygenation 

Tension-Based Indices 

P a 02 has withstood the test of time as an in- 
dicator of lung oxygenation function. A high P a o2 
on a low level of supplemental oxygen is nearly al- 
ways indicative of good lung oxygenation function. 
However, evaluation of P a o2 on increased F102 is 
somewhat more difficult. Three oxygen-tension- 
based indices attempt to address the problem of 
inteipretation of P a o2 when F102 is increased. 

Alveolar-arterial oxygen tension difference 

[PiA-aiOi]- The P(A- a i02- the difference between al- 
veolar and arterial P02 (or the alveolar-arterial oxy- 
gen tension gradient), was initially suggested as a 
sensitive screening tool to determine the presence 
of impaired lung oxygenation function. 4 Normally 
the P a o2 increases with increases in F102 (and, thus 
P102 and PAO2) (Fig. 1 ). As intrapulmonary shunt 
(Q sp /Q,) increases, the response of P a o2 to increases 
in F102 is severely attenuated (Figs. 1 & 2), but 
P a co2 changes very little (Fig. 3). The P ( A- a )02 in- 


180 - 


160 - 



140 - 


^120 - 


/ ) 

J 20% / 

1 100 - 

/ / 

80 - 

/ / 


* /S 30°/o/"*^ 

60 - 

~^^"^ 40%_____»- 

40 - 

. — • •~~~"*"~~50% 

1 1 1 1 



40 60 80 

% Inspired Oxygen 


Fig. 1. The effect of changing venous admixture (shunt) 
on P a o2 at different F102 levels. (Reprinted from Refer- 
ence 6, with permission.) 

creases at higher F102. and the increase in P(A-a)02 
is greater at higher levels of intrapulmonary shunt 
(Qsp/Qi) (Fig- 4). The P(A-a)02 is affected by many 
factors — including F102- ventilation/perfusion (V/Q) 
mismatch (Fig. 5), Q sp /Q t , and right-to-left intra- 
cardiac shunting. 4 ' 5 Furthermore. V/Q mismatch is 
affected by underlying lung disease, the age of the 
patient, the position of the patient (upright versus 
supine), and other factors common in critically ill 






\ 0.80\ 



\ \0.40 




10 20 30 40 50 

% Shunt 

Fig. 2. The effect of changing F102 on P a 02 at different 
degrees of intrapulmonary shunt. (Reprinted from Refer- 
ence 6, with permission.) 

patients. Finally, the P(A-a)0: ' s significantly al- 
tered by factors that change mixed venous oxygen 
content (CvCb)- These factors include tissue oxygen 
consumption (Vo 2 ), cardiac output (CO.), and he- 
moglobin concentration (tHb). 

20 30 40 50 
% Shunt 

Fig. 3. The effect of increasing shunt on P a o2 and P a C02. 
(Reprinted from Reference 6, with permission.) 

The P(A-a)0: takes into account the effects of 
changes in F102 and PaCO: xh and can be helpful in 
sorting out the magnitude of oxygenation dysfunc- 
tion that occurs in patients who are hypoventilating 
and have increased PaCO:- Although this index has 




F|0 2 

Fig. 4. The effect of increasing F102 on P(A-a)02 at differ- 
ent degrees of intrapulmonary shunting. (Reprinted from 
Reference 6, with permission.) 

% Inspired O2 

Fig. 5. The effect of changing F102 on P(A-a)02 at different 
degrees of V/Q mismatching. (Reprinted from Reference 
6, with permission.) 




proven to be a sensitive indicator of oxygenation 
dysfunction in patients breathing room air, it is quite 
nonspecific. At increased F102- the P(A-a)C>2 in- 
creases markedly, causing this index to lose clinical 
utility in critically ill patients. 

Table 1. Correlations between Measured Intrapulmonary 
Shunt (Qsp/Qt)* and Gas Exchange Indices 


r Value 

Estimated shunt 
Respiratory index, 

Arterial-alveolar oxygen fraction, 

Pa0 2 /PAO: 

Oxygenation index, 

Pa0 2 /Fl02 

Alveolar-arterial oxygen tension 
difference, P(A-ai02 






*Mean (standard deviation) of measured Qsp/Qt = 0.22 (0.1 1), 
with range = 0.33-0.53. Modified from Reference 15. with 

Because the P(A-a)02 does not correlate well with 
the degree of venous admixture present (Table 1, 
Fig. 6) and is affected by many nonpulmonary fac- 
tors, 7 it has lost favor with many critical care phy- 
sicians, may be of limited value in the critically ill 
patient, and for that reason is almost never used in 
our ICU. 

P(A-a)0 2 

Fig. 6. Poor correlation between Q sp /Qt and P(A-a)02- 
(Reprinted from Reference 35, with permission.) 

The respiratory index (RI). The respiratory index 
was devised to overcome some of the problems 
with the P(A-a)02 X ' 9 and ' s calculated as 


The RI is a more specific quantifier of oxygena- 
tion dysfunction of the lung than the P(A-a)02 and 
correlates somewhat better with the Qsp/Qt- 9 ' 10 

Arterial-alveolar oxygen tension ratio [P(a/A)02]- 

The arterial-alveolar oxygen tension ratio was like- 
wise devised to overcome some of the problems 
with the P(A-a)02 and is calculated as""' 3 

arterial-alveolar oxygen fraction = 

which may also be written as P(a/A)02- 


It, too, is a more specific quantifier of oxygena- 
tion dysfunction of the lung than the P(A-a)02< 
allowing comparison at different Ficks and estima- 
tion of P a 02 at a given F102 if Pa02 and P a c02 at an- 
other F102 are known, 14 and correlating somewhat 
better with the Q S p/Qt. 7 

Oxygenation index (OI). In 1974, Horovitz et al 15 
described the oxygenation index — an index used to 
compare arterial oxygenation at different F102S. 

01 = 



Cane et al 7 showed that the correlation between 
OI and Qsp/Qt is approximately the same as the cor- 
relation of RI, or P(A-a)02 / Pa02^ and the arterial- 
alveolar 2 tension ratio, or Pa02/PA02- with Qsp/Qt 
(Table 1 and Fig. 7), and clearly is more easily calcu- 
lated and, therefore, may have more clinical utility. 

•«&!■ ' 

: ': i «!!ii!iii:£ii : . 

Pa0 2 /Fl0 2 

Fig. 7. Relatively good correlation between Q sp /Qt and 
Pa02/Fi02- (Reprinted from Reference 35, with permis- 




A detailed review of the relationship between 
tension-based oxygenation indices and Qsp/Qt by 
Rasanen and Downs and co-workers 16 showed a 
good correlation between oxygen-tension-based 
variables and Qsp/Qt when oxygenation function is 
good (high P a o;) but a poor correlation in times of 
impaired lung function. Further, the indices do not 
take into account changes in peripheral extraction 
of oxygen that may lead to venous desaturation. 7 
Therefore, states associated with high Vo 2 or low 
CO. affect these indices. The presence of marked 
venous desaturation (when right-to-left shunting 
occurs) contributes to arterial hypoxemia. Greater 
shunting results in greater changes in arterial oxy- 
genation with venous desaturation. It should be 
clear that clinical conditions associated with ve- 
nous oxygen desaturation lower the PuC^/Pao: and 
increase P(A-a)02/PaO:- 

As previously indicated, two of the goals in the 
assessment of oxygenation are to determine the 
adequacy of tissue oxygenation and to determine 
which elements of the oxygenation dysfunction 
should be manipulated to improve the patient's 
overall course. Unfortunately, the tension-based in- 
dices described give no information regarding the 
adequacy of tissue oxygenation and do not provide 
information to guide therapeutic efforts beyond in- 
creasing Pio:- 

Saturation- and Content-Based Indices 

Although Qsp/ Qt estimations had been used for 
many years, indices based on oxyhemoglobin sat- 
uration and oxygen content were viewed with re- 
newed interest during the 1980s. The renewed in- 
terest came not so much because of a change in our 
understanding of the disease process but more be- 
cause of a change in technology. In the early and 
mid-1980s, developments in fiberoptic technology 
allowed assessment of both S a o: and mixed venous 
oxyhemoglobin saturation (SvO:) on a continuous, 
real-time basis, 17 from which both arterial (C a o 2 ) 
and mixed venous content (Cv0 2 ) could be deter- 

Oxyhemoglobin saturation can be assessed by 
noninvasive pulse oximetry, yielding an estimate of 
S a O: (designated as S p o 2 ). Continuous pulse oxim- 
etry allows the on-line assessment of functional oxy- 
hemoglobin saturation in many intensive care unit 

patients. At least in some settings, application of 
pulse oximetry has been responsible for a reduction 
in the measurement of arterial blood gases to assess 
the adequacy of arterial oxygenation. Is Pulse oxim- 
etry has an advantage over earlier attempts at con- 
tinuous oximetry because of the presence of less ar- 
tifact caused by changes in tissue and venous de- 
saturation in low-flow states. Although pulse oxim- 
etry is not always technically possible in low-flow 
states because of decreased pulsatility of the signal, 
it represents an accurate assessment of arterial sat- 
uration when it can be measured. Pulse oximetry is 
said to fail hard (that is to give no information) 
rather than fail soft (that is to give inaccurate infor- 
mation) during low-flow states. 

Pulse oximetry does not account for effects of 
carboxyhemoglobin, methemoglobin, or sulfhemo- 
globin. The presence (or absence) of dysfunctional 
hemoglobin species can be initially determined by 
CO-oximetry to avoid incorrect estimation of the 
actual S a 02- 

True mixed venous blood is found only after 
complete mixing of blood from the superior and in- 
ferior vena cavae and the coronary sinus and is 
available clinically from the right ventricle and pul- 
monary artery. Continuous mixed-venous oximetry 
became available in 1980 as a modification of the 
flow-directed pulmonary artery catheter commonly 
used in critically ill patients. 19 Both pulse oximetry 
and mixed-venous oximetry allow continuous as- 
sessment of oxygenation on a real-time basis. 20 

The volume of oxygen delivered from the left 
ventricle each minute (Do 2 ) is the product of C a o 2 
and CO. Substituting cardiac index (CI) for CO. 
provides indexing to body surface area — important 
because metabolic demands vary with body size. 
Oxygen consumption (Vo 2 ) is the volume of gas- 
eous oxygen actually used by the tissues each min- 
ute and is calculated from the Fick equation as the 
product of arterial-venous oxygen content differ- 
ence [Ca-viCb] and CO. As with Do 2 . substitution 
of CI provides indexing to body surface area. 

Once the technology was available for the con- 
tinuous on-line assessment of arterial and mixed 
venous saturations, a number of derived variables 
could be calculated. 21 Some of the most readily 
available are given here. 

Oxygen delivery index (ODD. The oxygen de- 
livery index is calculated as 




ODI = (Ca 02 )(CI). 

where CI is the cardiac index, or CO. divided by body 
surface area. 

Table 2 gives information regarding the volume 
of oxygen delivered from the left ventricle each 
minute. The value is indexed to body surface area 
so normal values can be obtained and compared 
with actual values measured in individual patients. 

Arterial-venous oxygen content difference. Knowl- 
edge of arterial and venous saturations allows the 
determination of the arterial-venous oxygen con- 
tent difference, C( a -v)02- 

This index is an important indicator of the rel- 
ative balance between CO. and Vo 2 . An increase in 
the C(a-v)02 indicates greater extraction of oxygen 
at the tissue level, and values above 5.5 mL/dL 
suggest that CO. is inadequate for the current level 
ofVo 2 . 

Oxygen consumption. The Fick principle states 
that the volume of oxygen consumed by the tissues 
is equal to the difference between the volume of 
oxygen delivered to the tissues and the volume of 
oxygen returned to the right side of the heart. The 
volume of oxygen delivered to the tissues is pre- 
sumed to be equal to Do2- The volume of oxygen 
returned to the heart is equal to the product of C v Cb 
and the venous return. Because over several heart- 
beats venous return and CO. must be equal, the to- 
tal volume of oxygen returned to the heart is equal 



V02 = C,a-v)02 X CO. 

Oxygen utilization coefficient (OUC). Although 
V02 is of interest in itself, the relationship between 
Vr> and D02 is of even greater interest. The OUC 
(sometimes called the : extraction ratio) is 

OUC = 


OUC can be estimated from S a 02 (via pulse ox- 
imetry) and Sv02 ( y i a venous oximetry): 

OUC = 

SaO: - SvO: 


This index describes the relative fraction of D02 
that is consumed by the tissues. Normally, only 
about 25% of delivered oxygen is consumed. Dur- 
ing times of stress when either consumption is in- 
creased or delivery is decreased, the OUC rises. 
OUCs greater than 0.35 imply an excessively high 
extraction of oxygen to meet the patient's current 
metabolic needs. Current definitions of the shock 
state often include references to the OUCs being 
greater than 0.35. 2223 

Mixed venous oxyhemoglobin saturation (SvOi)- 

S V 02 is important to the calculation of C v o:. How- 
ever, this value is also of interest in its own right. 24 
The Fick principle tells us that C v o: equals C02 
minus the volume of oxygen consumed by the tis- 
sues. Because virtually all of the venous oxygen 
content is bound to hemoglobin, the S v rj2 must be 
determined by the relationship between D02 and 
Vo 2 . In fact, the terms in the Fick equation can be 
rearranged to show that when S a 02 is maintained at 
a high level, the SvCb is mathematically ordained to 
be inversely related to the V02/D02 (Table 3 & Fig. 
8). Therefore, S V Q2 is an indication of the relative 

Table 2. Oxygen Transport Variables Commonly Used in the Assessment of Critically 111 Patients* 




Arterial content 
Venous content 
Oxygen delivery 
Oxygen consumption 
Oxygen utilization 

Terms are defined in the text. 


mL • min"' ■ i 
mL • min -1 ■ i 


C a o: = (Hb x SaO; x 1.34) + (P a O; x 0.0031) 

C v0: = (Hb x SvO: x 1.34) + (P v o : x 0.0031) 

Do: = C a O: x CI x 10 

Vo 2 = C(a-v)02 X CI xiO 

OUC = V (): /D>: 




balance between Vo : and Do: (thus, the OUC) 
can be approximated by 242 "' 


Sv02 - 



1 - OUC. 

The S\0; is indicative of the content of oxygen re- 
turned to the right side of the heart for reoxygena- 
tion. The four primary determinants of S V 02 are S a 02- 
hemoglobin concentration, CO., and Vo 2 . SvO: rep- 
resents the flow-weighted average of the venous ef- 
fluents from all perfused vascular beds. A low S v c>2 
indicates an imbalance between the Do: from the 
left ventricle and the consumption of oxygen by the 
tissues. 2 " It is a sensitive but nonspecific indicator 
of this imbalance. A normal or high SvO: does not 
necessarily indicate that the balance between D02 
and Vo: is adequate in all vascular beds. S v c>2 is af- 
fected more by low-extraction, high-flow tissues 
(such as the kidney) than by high-extraction, low- 
flow vascular beds (such as the myocardium). Sv02 
is not an indicator of tissue oxygenation but rather 
an indicator of the global balance between oxygen 
supply and demand. Although saturation- and con- 
tent-based oxygenation indices do not tell about tis- 
sue oxygenation per se, they do give us important 
information about the adequacy of global oxygen 
delivery and blood flow to the tissues. The overall 
balance between tissue Vo : and the D02 from the 
left ventricle may be assessed using the OUC. The 
relative balance between tissue Vo 2 and blood flow 
can be assessed by calculating C( a -v)02- These two 
indices give important information regarding the 
matching of Vo : with D02 and blood flow. 26 They 

Oximetrix Sv0 2 (%) 

Fig. 8. Correlation between Sv02 and oxygen utilization 
coefficient. (Reprinted from Reference 24, with per- 

give very little information about the adequacy of 
oxygen exchange in the lungs. 

Intrapulmonary shunt (Qsp/Qt). The calculation 
of Qsp/Qt gives an indication of venous admixture 
occurring at any point between the central venous 
circulation and the systemic arterial circulation. 
Qsp/Qt is an oxygen-content-derived variable. It is 
calculated from a simplified two-compartment 
model in which part of the blood reaching the left 
side of the circulation is completely oxygenated by 
passage through ventilated lung segments that are 
perfused, and part of the blood is unoxygenated be- 
cause it passes from the right side of the circulation 
to the left without exposure to ventilated alveoli. 
Venous admixture (or Qsp/Qt) is calculated as the 
difference between the content of fully oxygenated 
pulmonary capillary and arterial blood divided by 

Table 3. Derivation of Mixed Venous Oxygen Saturation (SvO;)* 

I Fick equation 
I solve for Q a _v)02 
I define for Qa-v)02 
I subtract C a o; 
I multiply by -1 
I divide by C a 02 
I define D02 
I define OUC 
I define S V 02 
I substitute S\02 

*Sv02 = mixed venous oxygen saturation. Vo; = oxygen consumption. Q,= total cardiac output, C (a -\)02 = arterial-venous oxygen con- 
tent difference, C a 02 = arterial oxygen content, C V 02 = mixed venous oxygen content difference, D02 = oxygen delivery, OUC = oxy- 
gen utilization coefficient. S a Q2 = arterial oxygen saturation. 

Step 1 



Qt X C (a -v)02 x 10 

Step 2 

C( a -v)02 


Vo;/(Q, x 10) 

Step 3 

Ca02 - C( a -v)02 


Vo 2 /(Q, x 10) 

Step 4 

- Cv02 


V 0: /(Q, x IO)-C a0: 

Step 5 



Ca02-Vo2/(Q, x 10) 

Step 6 

Cv02/C a 02 


1 -V 02 /(Ca02X Q, x 10) 

Step 7 

Cv0 2 /C a 02 


1 - V02/D02 

Step 8 

Cv02/C a 02 


1 -OUC 

(ifSa02 = 

1.0 (hen C 

02^C a 02 = Sv02> 

Step 9 



1 -OUC 




the difference between fully oxygenated pulmonary 
capillary blood and mixed venous blood: 

Q S p/Qt = 

C c '02 - C a o: 
Cc'O: - Cvo: 

Pulmonary capillary content (Cc'02) is estimated 
by calculating the Pao: assuming 100% saturation 
of pulmonary capillary hemoglobin. Pao2 is cal- 
culated as noted earlier by the alveolar gas equa- 

Cc'02 is then estimated as 

C c 'o: = (Hb x 1.34) + (0.0031 x P A0: ). 

C a 02 is calculated as 

C a o: = (Hb x Sao: x 1.34) + (0.0031 x P a02 ). 

Mixed venous oxygen content is calculated from 
a similar equation; but because dissolved oxygen 
adds so little to the total, it can usually be deleted 
or assumed to be in the middle of the physiologic 
range (35 torr). 

Cv02 = (Hb x Svo: x 1.34) + (0.0031 x Pv 02 ). 

The intrapulmonary shunt fraction, Qsp/Qt, is the 
gold standard used for the clinical assessment of 
lung oxygenation function. The utility of Qsp/Qt as 
an indicator of lung oxygenation function is not 
compromised by changes in Vo 2 , tHb, or Sv02- 
When calculated while the patient is breathing 
100% oxygen, the Q sp / Qt yields the fraction of 
blood passing from the right to the left side of the 
circulation through physiologic and anatomic 
shunts. When Qsp/Qt is measured on lower oxygen 
concentrations, it yields the fraction of blood pass- 
ing through physiologic and anatomic shunts as 
well as areas of severely low ventilation-perfusion 
ratio that are acting like physiologic shunts. The 
Qsp/Qt must be measured while the patient is 
breathing a known inspired oxygen fraction so that 
alveolar Po 2 can be calculated accurately. It also 
makes the assumption that pulmonary capillary 
blood is fully saturated and therefore requires an in- 
spired oxygen fraction of greater than 0.30 to as- 
sure complete pulmonary capillary saturation. The 
Qsp/Qt is the most accurate indicator of lung oxy- 

genation function that is clinically available at the 
patient's bedside. The calculation of Qsp/Qt there- 
fore requires the measurement of mixed venous ox- 
ygen saturation, and equations assuming (but not 
measuring) a normal mixed venous saturation or ar- 
terial-venous oxygen content difference should not 
be used in critically ill patients. 

Fractional arterial oxygen saturation (Fo^Hb). 

Fo2Hb is the best indicator of the C a 02 and is calcu- 
lated by dividing the oxyhemoglobin concentration 
(0 2 Hb in g/dL) by tHb (the sum of oxyhemoglobin, 
deoxyhemoglobin, carboxyhemoglobin, methemo- 
globin, and sulfhemoglobin). Oxygen bound to he- 
moglobin accounts for 95% or more of the total 
volume of oxygen in arterial blood (for hemoglobin 
levels of 10-15 g/dL and saturation > 90%). The 
P a 02 is directly related to the volume of oxygen dis- 
solved in the plasma, but dissolved oxygen contrib- 
utes very little to the total oxygen content. For this 
reason, I believe that arterial blood gas analysis for 
the assessment of oxygenation in critically ill pa- 
tients should play a minor role in therapy. Our in- 
tensive care unit, which formerly measured arterial 
blood gas values several times a day, now uses that 
analysis for the assessment of acid-base disorders 

Arterial-venous oxygen saturation difference 

[S(a-v)02]- S(a-v)02 is an indicator of the relative 
amount of oxygen extraction by the tissue. It re- 
flects the C(a-v>02 but can be measured continu- 
ously with combined arterial and venous oximetry. 

Ventilation-perfusion index (VQI). Ventilation- 
perfusion index (VQI) is a term coined by Riisanen 
and Downs in the mid-1980s. The VQI is an 
approximation of Qsp/Qt that can be performed on a 
continuous basis using combined arterial and 
mixed venous oximetry and can be calculated as 


1 - S a o: 
1 - Svo: 

The VQI reflects Qsp/Qt when the arterial ten- 
sion (and therefore saturation) is reduced. There- 
fore, when a high level of intrapulmonary shunting 
is present, and S a 02 is less than 1 .0. the VQI gives 
an accurate indication of Qsp/Qt. 27 "" 




Future Direction of Oxygenation Monitoring 

New technology is rapidly emerging that im- 
proves our ability to assess the patient's oxygena- 
tion status. Commercially available dual oximetry 
combining both arterial pulse oximetry and con- 
tinuous invasive mixed-venous oximetry allows a 
rapid assessment of oxygenation function and im- 
proved titration of therapy to physiologic end 
points. Continuous arterial blood gas analysis using 
fiberoptic technology will soon be clinically avail- 
able. 30 This should allow nearly complete elimina- 
tion of arterial blood gas analyses in the sickest of 
ICU patients; but this remains to be seen, and cost- 
effectiveness remains to be demonstrated. Measure- 
ment of subcutaneous tissue oxygen tension or in- 
tramuscular oxygen tension may soon be available. 31 

Other new technologies and resurrections of old 
technologies are improving our ability to detect 
failures to maintain tissue oxygenation. Accurate, 
clinically available lactic acid analyzers are now 
common in intensive care units and may detect the 
breakdown in oxygen transport that results in anae- 
robic metabolism and the accumulation of excess 
lactate. When the demand for oxygen exceeds the 
actual consumption of oxygen, tissues must switch 
to anaerobic pathways to achieve some energy pro- 
duction. |c, ' 3:; If anaerobic pathways are not avail- 
able, the intracellular energy crisis leads rapidly to 
death of the cell. The detection of excess lactate in 
arterial blood may indicate that anaerobic me- 
tabolism has taken place. However, because lactate 
can accumulate in poorly perfused peripheral tis- 
sues and later wash out with improved perfusion, 
an increased lactic acid level does not necessarily 
mean ongoing anaerobic metabolism. Furthermore, 
excess lactate must be metabolized (primarily by 
liver and kidneys) for the arterial levels to return to 
normal. An increased arterial lactate may indicate 
impaired breakdown rather than continued pro- 

Gastric tonometry offers promise for the assess- 
ment of the perfusion of the gastric mucosa. Most 
low-flow states cause an early redirection of blood 
flow away from the splanchnic organs. 33 - 34 Reduc- 
tion of mucosal blood flow decreases the removal 
of C0 2 and alters Pcc>2 accumulation in a semi- 
permeable intragastric balloon. Measurement of the 

Pcoa in the saline-filled balloon adjacent to the gas- 
tric mucosa allows calculation of mucosal inter- 
stitial pH (an estimate of the adequacy of gastric 
blood flow). It appears that gastric tonometry al- 
lows early detection of gastric mucosal hypoper- 
fusion and may be the first of several methods of 
evaluating regional or tissue-specific oxygen trans- 
port balance. 

In Conclusion 

A number of directly measured and derived oxy- 
genation indices are clinically available. 35 In the 
last 5 to 10 years there has been a move away from 
oxygen-tension-based indices to saturation- and 
content-based indices. These content-based indices 
allow the assessment of lung function and the over- 
all balance between oxygen delivery and oxygen 
consumption. The goal in the care of critically ill 
patients is to maintain tissue oxygen delivery so 
that oxygen consumption can increase to meet the 
metabolic requirements of the tissues. Direct as- 
sessment of tissue oxygenation is not available in 
the early 1990s, but the assessment of global oxy- 
gen transport balance is clinically available and 
may improve the care of critically ill patients. 


I thank Ms Selma Archer for the preparation of the manu- 
script and Ms Karen Safcsak RN for her careful review and 


1. Shapiro BA, Harrison RA. Cane RD, Kozlowski- 
Templin R. Arterial oxygenation. In: Clinical applica- 
tion of blood gases, 4th ed. Chicago: Year Book Med- 
ical Publishers, 1989:57-76. 

2. Cole PV. Bench analysis of blood gases. In: Ledingham 
IM. ed. Respiratory monitoring in intensive care. Edin- 
burgh: Churchill Livingstone. 1982:32-55. 

3. Nelson LD, Snyder JV. Technical problems in data ac- 
quisition. In: Snyder JV, Pinsky M. eds. Oxygen trans- 
port in the critically ill. Chicago: Year Book Medical 
Publishers, 1987:205-234. 

4. Neff TA. Monitoring alveolar ventilation and res- 
piratory gas exchange. RespirCare 1985:30:413-421. 

5. Viale JP, Percival CJ, Annat G. Rousselet B, Motm I. 
Arterial-alveolar oxygen partial pressure ratio: a theo- 
retical reappraisal. Crit Care Med 1986:14:153-154. 




6. Dantzker DR. Pulmonary gas exchange. In: Dantzker 
DR. ed. Cardiopulmonary critical care. 2nd ed. Phil- 
adelphia: WB Saunders Co, 1991:25-44. 

7. Cane RD. Shapiro BA. Templin R, Walther K. Un- 
reliability of oxygen tension-based indices in reflecting 
intrapulmonary shunting in critically ill patients. Crit 
Care Med 1988:16:1243-1245. 

8. Shapiro BA. Arterial blood gas monitoring. Crit Care 
Clin 1988:4:479-492. 

9. Siegel JH, Farrell EJ. Miller M. Goldwyn RM. Fried- 
man HP. Cardiorespiratory interactions as determinants 
of survival and the need for respiratory support in hu- 
man shock states. J Trauma 1 973 ; 1 3 :602-6 1 9. 

10. Sganga G. Siegel JH, Coleman B, Giovannini I, Bol- 
drini G, Pittiruti M. The physiologic meaning of the res- 
piratory index in various types of critical illness. Circ 
Shock 1985;17:179-193. 

11. Hess D. Maxwell C. Which is the best index of oxyge- 
nation — P(A-a)02. PuOi/Pao:. or PaO:/Fio:? (editorial). 
RespirCare 1985;30:961-963. 

12. Gilbert R, Auchincloss JH Jr, Kuppinger M, Thomas 
MV. Stability of the arterial/alveolar oxygen partial 
pressure ratio — effects of low ventilation/perfusion re- 
gions. Crit Care Med 1979:7:267-272. 

13. Peris LV, Boix JH. Salom JV, Valentin V, Garcia D. 
Arnau A. Clinical use of the arterial/alveolar oxygen 
tension ratio. Crit Care Med 1983:1 1:888-891. 

14. Gilbert R, Keighley JF. The arterial/alveolar oxygen 
tension ratio. An index of gas exchange applicable to 
varying inspired oxygen concentrations. Am Rev Respir 
Dis 1974;109:142-145. 

15. Horovitz JH, Carrico CJ, Shires GT. Pulmonary re- 
sponse to major injury. Arch Surg 1974;108:349-355. 

16. Rasanen J, Downs JB, Malec DJ, Oates K. Oxygen ten- 
sions and oxyhemoglobin saturations in the assessment 
of pulmonary gas exchange. Crit Care Med 1987:15: 

17. Baele PL, McMichan JC, Marsh HM. Sill JC, Southorn 
PA. Continuous monitoring of mixed venous oxygen 
saturation in critically ill patients. Anesth Analg 1982; 

18. Bierman MI, Stein KL, Snyder JV. Pulse oximetry in 
the postoperative care of cardiac surgical patients — a 
randomized controlled trial. Chest 1992;102:1367-1370. 

19. Kandel G, Aberman A. Mixed venous oxygen satura- 
tion: its role in the assessment of the critically ill pa- 
tient. Arch Intern Med 1983:143:1400-1402. 

20. Nelson LD. Monitoring lung injury. In: Critical care: 
state of the art. Fullerton CA: Society of Critical Care 
Medicine. 1987:245-278. 

21. Nelson LD. Monitoring oxygen delivery. Proc R Coll 
Physicians (Edinb) 1991:21:401-408. 

22. Nelson LD. The state of shock. Acute Care Therapeutics 

23. Nelson LD. Monitoring and measurement in shock. In: 
Barrett J. Nyhus LM. eds. Treatment of shock. 2nd ed. 
Philadelphia: Lea and Febiger, 1986. 

24. Nelson LD. Continuous venous oximetry in surgical pa- 
tients. Ann Surg 1986;203:329-333. 

25. Rasanen J. Downs JB, Malec DJ. DeHaven B. Seidman 
P. Estimation of oxygen utilization by dual oximetry. 
Ann Surg 1987:206:621-623. 

26. Nelson LD. Rutherford EJ. Monitoring mixed venous 
oxygen. RespirCare 1992:37:154-164. 

27. Rasanen J, Downs JB, DeHaven B. Titration of continu- 
ous positive airway pressure by real-time dual oximetry. 
Chest 1987:92:853-856. 

28. Carroll GC. A continuous monitoring technique for 
management of acute pulmonary failure. Chest 

29. Bongard FS. Leighton TA. Continuous dual oximetry in 
surgical critical care: indications and limitations. Ann 
Surg 1992;216:60-68. 

30. Shapiro BA. Harrison RA. Cane RD. Kozlowski- 
Templin R. Fluorescent optode technology. In: Clinical 
application of blood gases. 4th ed. Chicago: Year Book 
Medical Publishers, 1989:307-312. 

3 1 . Hunt TK. Halliday BJ. Hopf HW, Scheuenstahl H. West 
JM. Measurement and control of tissue oxygen tension 
in surgical patients. In: Gutierrez G. Vincent JL, eds. 
Update in intensive care and emergency medicine: tis- 
sue oxygenation. Berlin: Springer- Verlag. 1991:337- 

32. Kasnitz P, Druger GL, Yorra F. Simmons DH. Mixed 
venous oxygen tension and hyperlactatemia: survival in 
severe cardiopulmonary disease. JAMA 1976:236:570- 

33. Fiddian-Green RG. Tonometry: theory and applications. 
Intensive Care World 1992:9:61-65. 

34. Hess D. Kacmarek RM. Techniques and devices for 
monitoring oxygenation. RespirCare 1993:38:646-671. 

35. Covelli HD. Nessan VJ, Tuttle WK 3rd. Oxygen de- 
rived variables in acute respiratory failure. Crit Care 
Med 1983.1 1:646-649. 




Nelson Discussion 

Dantzker: I'd like to challenge a 
couple of basic concepts that came 
up this morning. I think we have to 
challenge the concept that you have 
only two choices — acidosis or death. 
I'm not sure that those are the only 
two choices. I don't know that I'm 
going to give you others right now, 
but I think that the system is much 
more complex. For example, I can 
remember some studies that Guillier- 
mo Gutierrez did using magnetic res- 
onance spectroscopy (MRS). 1 He re- 
duced oxygen transport by ischemia 
and by hypoxemia, got to a point 
where the high-energy phosphate 
compounds were reduced by the 
same degree at the end of the period 
of stress, and then reoxygenated 
them. The tissues that had been made 
ischemic resynthesized high-energy 
phosphates and came back to look 
very much the same as they had be- 
fore. The hypoxemic tissues con- 
tinued to drop their high-energy 
phosphates, their ATP began to fall, 
and there was something very differ- 
ent about those tissues. I think the 
first tissue "did something' besides 
acidosis or death. I mean I think 
there are maybe very complex pat- 
terns. I recently saw a woman with 
Osler-Weber-Rendu syndrome who 
was a Christian Scientist. She bled 
continuously from her multiple gas- 
trointestinal a-v malformations and 
she had a Po: of 26 because of the a-v 
malformations, in her lung. When I 
saw her, her hemoglobin was 6 
grams. Now. if you calculate some of 
your indices, this lady should have 
been dead. However, she was sitting 
up, in bed, looking fine. I daresay I 
wouldn't want to know what her 
mixed venous oxygen tension and a-v 
O: difference were, or even less, 
what her coronary sinus oxygen ten- 
sion must have been. It must have 
been unmeasurable, and she was not 
acidotic. She was able to walk 
around. I mean, she obviously wasn't 

running any races, but I think there's 
more to it. I think it is dangerous for 
us to enshrine this simple concept as 
being the way to look at these pa- 
tients — because I'm not sure that's 
really what happens. Certainly, not 
on a chronic basis. Acutely, things 
may be very different. 

1. Gutierrez G, Pohil RJ, Andry JM, 
Strong R, Narayana P. Bioenergeties 
of rabbit skeletal muscle during hy- 
poxemia and ischemia. J Appl Physiol 

Nelson: What was the lady's cardiac 

Dantzker: 1 don't know. 

Wood: 14? 

Dantzker: Even if it was 14, with a 
Po; in the 20s, and a hemoglobin of 6 
grams, that's a wide a-v 2 differ- 
ence. That's a wide difference, even 
if it was 14. So, I'm just not sure that 
those are the two choices. 

Cerra: I have to agree with David 
(Dantzker) — but I think it's a mis- 
take to assume you're either in an- 
aerobic or aerobic metabolism. It's a 
continuum of response. Then, I think 
we have to put in another group of 
experiments done in reversible myo- 
cardial anoxia, which are best de- 
scribed as prolonged periods of ab- 
solute anoxia, with remarkable pre- 
servation of morphology on electron- 
microscopy until re-perfusion oc- 
curs.' We have to introduce this con- 
cept of the reperfusion injury and the 
subsequent production of the in- 
flammatory modulators of metabo- 
lism. I think that's what you're driv- 
ing at. 

1. Cerra FB. Lajos TZ, Montes M. Sie- 
gel JH. Structural-functional corre- 
lates of reversible myocardial anoxia. 
J Surg Res 1974:16:140-152. 

Dantzker: Well, either that or some 
change in regulation. There's lots of 
room for the cell to alter and evolve 

pathways and shunts — things it can 
do to try to down-regulate its energy 
requirements in order to maintain 

Cerra: Which indeed happens. The 
chronic cardiogenic patient is prob- 
ably a model for that. 

Nelson: I think you're exactly right. 
I think that must be what happens. 
There's a tremendous change of con- 
sumption that is not a bad change of 
consumption. I still am not sure you 
can convince me that if oxygen de- 
mand remains extraordinarily high 
and, in fact, cells are not consuming 
oxygen to meet demand that those 
cells don't have to switch to an- 
aerobic processes. I didn't mean to 
imply the whole body at all, but rath- 
er tissues that have not increased oxy- 
gen consumption to meet their aero- 
bic demands. 

Cerra: I didn't say that — but on one 
end of the spectrum is pure aerobic 
metabolism and on the other end of 
the spectrum pure anaerobic metabo- 
lism. And in between there are all 
combinations of electron transfer. 

Dantzker: Don't forget the turtles, 
you know. They obviously don't 
switch to acidosis. They just down- 

Cerra: They just down-regulate, ie, 

Dantzker: The same thing happens 
with animals. Do something to a dog. 
make him sick, and he'll just go off 
in a corner and not do anything. I 
mean, animals down-regulate. I think 
we have to be very careful about 
how we look at the system. 

Morris: I liked the presentation be- 
cause it fits with what I teach. 

Nelson: You taught me all this stuff, 
remember, when I was in Salt Lake. 

Morris: My concern is severalfold. 
First, we don't know anything about 
the distributions of blood flow and 




local oxygen demand. With the wide 
variety of oxygen extractions, patient 
conditions, and tissue oxygen levels 
for different organs (heart and kid- 
ney, to give two examples), it be- 
comes very difficult to know exactly 
what to do with decreases in mixed 
venous saturation, without knowing 
regional distributions. Second, you 
can probably help me here. Frank 
(Cerra) — if my recollection is cor- 
rect, isolated mitochondrial prepara- 
tions begin to decrease oxygen up- 
take when the Po: in equilibrium 
with the preparation varies some- 
where between 0.1 and 2 mm Hg. 

Cerra: That's right. 

Morris: It varies depending upon the 
experiment. That means, of course, 
that every specimen we have avail- 
able to us for sampling is rather far 
upstream from the site of oxygen 
utilization. That, in and of itself, 
should make us rather cautious in 
drawing any definitive conclusions 
about any of these measurements. I 
mean, a mixed venous Po: of 40 or 
20 might or might not reflect some- 
thing noxious at the system level. 
Just an anecdotal comment, which 
I've never understood, and I think 
it's equal to Dave's (Dantzker) — I 
took care of a lady years ago who 
had an arterial Po: of 17 to 19, in the 
throes of a terrible hypoxic res- 
piratory failure, mixed venous P02 
that varied from 11 or 12 to 14, con- 
sistently. She had segmentation of 
the retinal veins, which is a finding 
we use (for those of you not familiar 
with it) for people who are essential- 
ly dead, and she was responsive and 
alert. She had fixed, dilated pupils — 
unresponsive pupils, directly and 
consistently unresponsive — yet, she 
was quite alert. We talked about her 
daughter and having her daughter 
come in to visit her and tiny details. 
It left me with rather grave concern 
for some professors at Yale who 
taught me as a medical student that 
an arterial Po^ of 40 was incom- 

patible with life. Anyway, I'm con- 
cerned, as Dave (Dantzker) is. about 
fixing these concepts, although I 
think they provide a useful scheme 
for thinking about things. 

Nelson: With mixed venous oxygen 
saturation, like all the other indices 
that we talked of this morning, the 
normal, or better than normal, some- 
times, is difficult to interpret without 
other clinical information. Certainly 
when the mixed venous oxygen sat- 
uration is low, it should trigger some 
kind of response to investigate why 
it's low. In regard to the regional 
blood flow, I think everybody here 
recognizes very clearly that mixed 
venous oxygen saturation is a flow- 
weighted average of the effluents 
from the perfused vascular beds — 
and all those words are operative — 
because the beds have to be per- 
fused. If there's no perfusion, there's 
no impact, just as there's no impact 
from lactate that's being produced 
and stored someplace in a nonper- 
fused bed, but is not coming back to 
where we can measure it. The two 
beds that you mentioned, I think, are 
two good examples. Look at myo- 
cardium and look at kidney — the 
change in mixed venous oxygen sat- 
uration caused by a change in renal 
blood flow (that is. a high-flow, low- 
extraction bed) has a much greater 
effect on SvO; than does the change 
in a high-extraction, low-flow bed. 
like myocardium. Therefore, the cru- 
cial bed — already maximally extract- 
ing — has very little effect on SvO:- I 
think that is extremely important. 

Wood: With regard to the various in- 
dices of pulmonary oxygenation, I 
would really encourage avoidance of 
nosology. We have lot of possible 
descriptors of O: exchange to ap- 
proximate shunt, but then we just 
waste effort comparing one with an- 
other to see which is closest to shunt, 
and end up confused. The most tell- 
ing slide was the variability that you 

presented showing that shunt was the 
most reproducible. Seems to me that 
if you want to know what the lung 
O: exchange is in a patient with very 
bad air-space filling or other lung 
disease with shunting that you should 
measure the shunt. If you don't 
measure the shunt, everything else 
has so many associated caveats and 
assumptions that you could thought- 
fully have guessed — you have a wide 
range of error. Furthermore, your 
thoughtful best guess has less error 
when it is based on the measured car- 
diac output and the measured mixed 
venous O; content. Better you should 
just measure the shunt — or if you 
don't measure it, admitting that you 
have to come at lung O: exchange 
defects from other indirect ways 
when how much of the lung in the x- 
ray is flooded is as good as anything 
else. Regarding the indices of oxy- 
gen transport, it seems to me that 
measuring cardiac output and the ar- 
terial content tells you all that you 
need to know about the transport. 
Measuring the cardiac output and the 
a-v O; content difference tells you all 
you need to know about the con- 
sumption, especially if you add to it 
an estimate of excess lactate. Those 
are really the only two variables. I 
disagree with your interpretation that 
Terry Phang's paper made a dis- 
tinction between Fick-measured oxy- 
gen consumption and what the real 
oxygen consumption is by cart.' I in- 
terpreted his data to say when you 
make a variable of thermodilution 
cardiac output times the arterial con- 
tent and plot that against thermodi- 
lution cardiac output times a-v O; 
content difference, you can introduce 
a shared error, which accounts for a 
spurious relationship between Vo; 
and D02 that can be avoided by cart 
M);. But cart and Fick Vq : each meas- 
ure tissue O; consumption. 

The last comment that I want to 
make relates to hibernation. In fair- 
ness to you, what you had to do was 
present a variety of indices of lung 




and tissue oxygenation. That was the 
nature of your talk, and I'm grateful 
for how well you did that. This dis- 
cussion of your paper invites us to be 
careful about the assumptions under- 
lying supply-dependence of Vo: be- 
cause one of those assumptions is 
that you either have aerobic or an- 
aerobic metabolism, but our dis- 
cussion also considers reduced aero- 
bic metabolism and hibernation. For 
example, can a beaker of cells, when 
exposed to progressive, slow hypox- 
ia hibernate? If their oxygen con- 
sumption actually goes down when 
you do that, then there are data at a 
very cellular level that support the 
notion that maybe there's a "hi- 
bernin.' some kind of enzyme turn- 
on that down-regulates metabolism. 
That possible outcome would take 
the mixed venous saturation and 
question it a fair bit. It might be that 
when the Vo; falls off, as in circum- 
stances of unresuscitated shock, that 
what you actually have is hibernation 
going on. The cells are actually 
down-regulating their metabolism. 
At the 1992 ATS meeting, Schu- 
macker showed that there is a down- 
regulation of the Vo: of hepatic cells 
exposed slowly to hypoxia.' Fur- 
thermore, that complicates the whole 
business of when you left-shift the 
oxyhemoglobin dissociation curve, 
the critical Po: falls from 25 to 15. 2 
Wait a minute! What's doing that? If 
you go down from a Po: of 25 
(where cells usually get anaerobic or 
hibernate, whichever it is) all the 
way down to 15, with no change in 
the extraction, and you maintain the 
Vo: all the way down to that level, 
it's as if mixed venous Po: has noth- 
ing to do with this. It just happens 
that the Po? of the mixed venous 
blood was what happened when Vo 2 
began to decrease; so, in both cases 
you need a 50% shunt in the pe- 
riphery. Whether PvO : ) was 25 or 15, 
a 60-70% shunt was required to ac- 
count for these data. 2 In other words, 
if the cells at the mitochondrial level 

are 1 torr, you need for 60-70% of 
arterial blood lung not to lose any 
oxygen at all while perfusing tissues 
in which Vo: is decreasing in order to 
get a mixed saturation of 30-40%. 2 
So. where those shunts are is still un- 

1. Schumacker PT, Chande! N, Agusti 
AGN. Oxygen conformance of cellu- 
lar performance in hepatocytes. Am J 
Physiol Lung Cell Molec Biol 1993 
(in press). 

2. Schumacker PT. Long GR, Wood 
LDH. Tissue oxygen extraction dur- 
ing hypovolemia: role of hemoglobin 
P 50 . J Appl Physiol 1987;62:1801- 

Nelson: I agree with you. If you 
want to know about lung oxygena- 
tion function, you need the shunt. I 
would go even further than you did. 
The A-a gradient in critically ill pa- 
tients in the ICU is of no value what- 
soever. It's a screening tool for the 
clinic or the pulmonary function lab. 
It's not a tool for the ICU. However I 
would disagree with your comment 
about 'real' oxygen consumption. 
The real oxygen consumption is cal- 
culated by the Fick equation. That's 
thermodynamics; that is exactly how 
much oxygen those cells are using, 
and that is their consumption. What 
we measure by the cart is oxygen up- 
take, and that's varied by a whole 
number of things that are beyond 
what we have time to discuss today. 

Dantzker: Just to make things a lit- 
tle more complicated — Larry (Wood), 
of course, knows that even meas- 
uring the shunt doesn't necessarily 
tell us what the oxygenating ef- 
ficiency of the lung is. Both of us 
have done some work looking at the 
effects on the shunt of altering flow 
through the lungs. The shunt is al- 
tered as flow through the lung chang- 
es. So, as cardiac output goes up. for 
the same kind of abnormality in the 
lung, the level of shunt that's meas- 
ured increases. I showed that experi- 

mentally. 1 and I remember that Dr 
Lemaire looked at patients with 
ARDS who were treated with the ex- 
tracorporeal membrane oxygenator. 2 
He showed that as he dialed up the 
amount of flow that was going 
through the lungs, he could alter 
Qsp/Qi. So even the amount of shunt 
is not an absolute measure of the 
amount of gas exchange abnormality 
seen in the lung. I don't think any of 
us know why. I gave up — before 
Larry (Wood) — trying to find a rea- 
son. Maybe he wants to comment on 
it. Let me make one other quick com- 
ment. Schumacker' s studies were 
very interesting. 3 What he actually 
showed was that it depends on the 
hypoxia history of the cells as to 
what the critical point is. If you make 
them hypoxic slowly, you can shift 
the point at which that critical oxy- 
gen transport is reached. I thought 
that was an intriguing study, the first 
one I've seen trying to do that. So 
the cells are making some kind of 

Finally, let me just raise the topic 
of exercise. During exercise, there's 
an increase in O: demand that occurs 
immediately. Thus, during exercise, 
the widening of the a-v O; difference 
isn't bad. It's built right into the sys- 
tem that you should increase your 
extraction. And the increase in ex- 
traction increases to 80% or more at 
peak exercise. The same thing could 
be said about lactate. Lactate goes 
up, and for years we've all assumed, 
simplistically. that that means there 
are areas of muscle that are anae- 
robic. But even that's not so clear. 
Many exercise physiologists have 
debated for years about the cause of 
the anaerobic threshold. 4 Does it real- 
ly indicate the onset of anaerobic me- 
tabolism or is it just a change in the 
way in which muscle is handling glu- 
cose? We all know that there are lots 
of things that can lead to elevation of 
lactate, without involving anaerobio- 
sis. Anything that increases glucose 
transport or glucose utilization in ex- 




cess of the ability to go through the 
Krebs cycle will raise lactate levels. 

1. Lynch JP, Mhyre JG, Dantzker DR. 
Influence of cardiac output on intra- 
pulmonary shunt. J Appl Physiol: 
Respirat Environ Exercise Physiol 

2. Lemaire F, Jardin F. Harari A, et al. 
Assessment of gas exchange during 
venoarterial bypass using the mem- 
brane lung. In: Zapol WM. Qvist J, 
eds. Artificial lungs for acute res- 
piratory failure. New York: Academic 
Press, 1976. 

3. Schumacker PT, Chandel N. Agusti 
AGN. Decreased oxygen consump- 
tion in response to chronic hypoxia in 
isolated hepatocytes (abstract). Am 
Rev Respir Dis 1992:145(4. Part 2): 

4. Wasserman K. Anaerobiosis, lactate, 
and gas exchange during exercise: the 
issues. Fed Proc 1986:45:2904-2909. 

Nelson: I don't want to get into the 
therapy side of things, but I still have 
problems with the lung blood-flow 
effects on the calculated shunt. We 
did some studies, 1 stimulated by your 
studies, 2 a number of years ago in an 
animal model of severe ARDS (the 
oleic-acid model in dogs, which is 
not particularly good) and looked at 
volume-resuscitated and under-resus- 
citated animals treated with enough 
PEEP to maintain oxygen consump- 
tion and maintain oxygen delivery. 
What we found was that shunt goes 
up dramatically as the ARDS gets 
worse. As you go up on PEEP the 
shunt comes down, but in the un- 
resuscitated animals cardiac output 
fell tremendously. In the resuscitated 
animals, cardiac output was main- 
tained. In fact, shunt still came 
down, even though flow was main- 

1. Nelson LD, Houtchens BA, Westen- 
skow DR. Oxygen consumption and 
optimum PEEP in acute respiratory 
failure. Crit Care Med 1982:10:857- 

2. Danek SJ, Lynch JP. Weg JG. Dantz- 
ker DR. The dependence of oxygen 
uptake on oxygen delivery in the 
adult respiratory distress syndrome. 
Am Rev Respir Dis 1 980; 122:387- 

Dantzker: Oh. I'm not trying to tell 
you that falling cardiac output is the 
way that PEEP leads to improve- 
ment, but I'm saying that if you have 
a given level of shunt, if you dial that 
volume output up or down, you alter 
the level of shunt. Therefore, that's 
not an absolute measure, either. 

Morris: First, thank you for doing 
this presentation. I'm glad you're up 
and not me. I think you did what you 
were supposed to do. present oxy- 
genation indices. It seems to me, in 
response to your presentation and to 
Dr Wood's comment, that it would 
be better to under-measure shunt. 
One has to recognize that implicit in 
the statement by him that it would be 
better to measure something — that 
there's a goal. A lot of what we're 
talking about depends upon the goal. 
If our goal is to define the behavior 
of the lung, then it may be valuable 
to measure shunt and it may be ap- 
propriate, as you suggested, not to 
pay any attention to the pressure 
measurements, ratios, differences of 
arterial Po;. directly. If the goal, on 
the other hand, is to enhance survival 
or manage care, those things are not 
quite so clear. For example, taking 
care of a patient and helping a pa- 
tient recover and survive (either be- 
cause of or in spite of us) and return 
home may involve a completely dif- 
ferent set of imperatives than those 
that would direct us if we were inter- 
ested in defining the physiologic be- 
havior of an organ like the lung. To 
be quite frank, the patient with 
ARDS is probably very little inter- 
ested in his or her mixed venous sat- 
uration, Po:. A-a gradient, or shunt 
but quite interested in how much it 
hurts and whether they'll be able to 
breathe again and go back home. I 

think I'll come back to this tomorrow 
and perhaps raise some interesting 
foils to give you an opportunin to 
get back at me. I think that we have 
to be certain what our goals are. In 
my opinion, the reductionist thrust, 
which has characterized medical re- 
search since the end of World War 
II, has produced some very inter- 
esting and even titillating informa- 
tion about physiology, biochemistry, 
and so forth, but frequently has pro- 
duced rather little that has impacted 
patient care. I'll come back to this 

Cerra: I have just one comment that 
Larry (Wood) started and Dave 
(Dantzker) had asked about. It has to 
do with the type of fuel oxidized to 
make energy, high-energy phos- 
phate, and the effect of that fuel type 
on the redox state of the cytosol and 
mitochondria. This is part of the non- 
anaerobic metabolic compensatory 
mechanism? We're accustomed to 
focusing on glucose as the metabolic 
fuel. In fact, these cells, under in- 
flammatory conditions, aren't using 
only glucose. They're burning fat 
and they're burning protein. That has 
a couple of characteristics that are 
important for us to understand. You 
can have an elevated lactate-pyruvate 
ratio indicating a reduced cytosol re- 
dox state, normal mitochondrial 
readouts, and very normal fat oxida- 
tion, and not be in anaerobic me- 
tabolism at all. In this transition me- 
tabolism, the tangible expression of 
that is that oxidizing a mole of glu- 
cose gives you 3.4 calories, a mole 
of fat gives you anywhere from 7 to 
10 calories, depending on what 
chain-length of fat you're talking 
about. Short-chain fats are incredibly 
efficient oxidative fuels. They take 
very few moles of ATP to prepare, 
and they produce a number of moles 
of ATP and oxidation. That's part of 
what we need to fit into this. The 
problem is. to get back to Dr Mor- 
ris's point, what surrogate you would 
choose to measure? 




My concern is this question Loren 
(Nelson) raised. Does the uptake of 
oxygen from expired gas really 
measure oxygen consumption (ie, 
what are the factors that affect ex- 
pired gas analysis)? I'm not ques- 
tioning that the sensors are accurate. 
In a closed system the sensors are in- 
credibly accurate, even if you check 
them out with mass spectrographs as 
a gold standard. The problem is 
when you put that system on the ven- 
tilated patient, the error rate is very 
large, especially when the oxygen 
concentration is above 40 or 45%, 
because of the characteristics of the 
system (not the sensor) and leaks in 
the system. I think this is something 
we need to take up, in order to bring 
that piece back into focus. 

Nelson: I think Terry (Phang) 
showed with his data this morning 
that the system is incredibly ac- 

Cerra: In vitro. 

Nelson: In vitro. That's right. But. 
again, that was bleeding nitrogen and 
CO: into a black box, and you're 
right. When you put it on the patient. 

it's different. There are a number of 
sources of error. We've now — and 
you and I've talked about this be- 
fore — we have now a series of al- 
most 2,500 detailed metabolic meas- 
urements that we've correlated with 
cardiac outputs, and I can tell you, 
unequivocally, oxygen uptake and 
oxygen consumption are different. 
They're different things (unpub- 
lished data). 

Wood: I just want to say to Loren 
(Nelson) one more time that attrib- 
uting thermodynamics to the reason 
why Fick equation is really the oxy- 
gen consumption, may overlook the 
fact that it's equally thermodynamic 
to get the difference between oxygen 
inspired and oxygen expired. Though 
there are errors, and Frank (Cerra) 
was talking about them, between in- 
spired and expired oxygen fractions, 
they are smaller than the + 15% er- 
rors in thermodilution cardiac output. 
Terry (Phang) spent a lot of time this 
morning showing us his details. I 
think everybody in this room under- 
stands that there are intrinsic errors 
in thermodilution cardiac output, ar- 
terial oxygen content, and, especial- 

ly, mixed venous oxygen content, 
and that the difference between those 
two O: content measures multiplied 
by thermodilution cardiac output is 
much more likely to be the source of 
error in the measurement of Vo : than 
that caused by measures of minute 
ventilation and the difference be- 
tween Fio; and Feo; corrected for 
changing nitrogen. 

Nelson: But the errors in thermo- 
dilution cardiac output are trivial 
compared to the errors in the uptake 

Wood: No way. 

Nelson: Trivial! 

Morris: There was no mention this 
morning of the difference between 
gas-phase oxygen uptake measure- 
ments and blood-phase oxygen up- 
take measurements (a result of oxy- 
gen consumption by the lung). I 
would just remind everybody that the 
lung can account for 20-40% of the 
total metabolic rate of a paralyzed 
patient. So, you can't assume that 
differences between the two are er- 



Techniques and Devices for Monitoring Oxygenation 

Dean Hess MEd RRT and Robert M Kacmarek PhD RRT 

I. Introduction 

II. Arterial Blood Gas Analysis 

III. Point-of-Care Testing 

IV. Continuous In-Vivo Evaluation of Arterial Oxygenation 
V. Pulse Oximetry 

VI. Transcutaneous P02 

VII. Mixed Venous Blood Gas Assessment 

VIII. Venous Oximetry 

IX. Dual Oximetry 

XI. Oxygen Consumption 

XII. Continuous Fick Cardiac Output 

XIII. Gastric Tonometry 

XIV. Monitoring in Perspective 


The real and potential effects of hypoxia are fa- 
miliar to all clinicians who practice critical care. 
Because of this fear of hypoxia and its effects on 
homeostatic balance, it is not surprising that many 
technologies have been developed to measure and 
monitor oxygenation. Some of these have become 

Mr Hess was Assistant Director, Department of Research, 
York Hospital, and Instructor. School of Respiratory Therapy. 
York Hospital and York College of Pennsylvania. York. Penn- 
sylvania, when this paper was prepared: he is now Assistant 
Director, Respiratory Care, Massachusetts General Hospital, 
and Instructor. Department of Anesthesia, Harvard Medical 
School — Boston, Massachusetts. Dr Kacmarek is Director, 
Respiratory Care, Massachusetts General Hospital, and Assist- 
ant Professor, Department of Anesthesia, Harvard Medical 
School, Boston, Massachusetts. 

A version of this paper was presented by Mr Hess on October 
8, 1992, during the Respiratory Care Journal Conference on 
Oxygenation in the Critically 111 Patient, held in Puerto Val- 
larta, Mexico. 

The authors have no financial interest in the products men- 
tioned or in competing products. 

Reprints: Dean Hess MEd RRT, Ellison-4. Massachusetts Gen- 
eral Hospital, 32 Fruit St, Boston MA 02 1 14. 

ubiquitous within the past 10 years (eg, pulse oxim- 
etry). Devices to evaluate oxygenation can be cat- 
egorized in several ways (Fig. 1 ). These devices 
may be invasive or noninvasive, continuously or in- 
termittently operative. 1 " 5 They can be used to evalu- 
ate arterial oxygenation, venous oxygenation, oxy- 
gen delivery, oxygen uptake, or a combination of 




Fig.1 Characteristics of oxygenation evaluation tech- 

Arterial Blood Gas Analysis 

Arterial blood gas (ABG) analysis is the tradi- 
tional method and, indisputably, the gold standard, 
for evaluation of arterial oxygenation. Electro- 
chemical (electrode) techniques to measure the par- 
tial pressure of oxygen (P02) ' n whole blood were 
introduced by Dr Leland Clark in the mid-1950s." 




By the mid-1970s, 24-hour availability of blood gas 
measurements had become possible in most critical 
care units. Although guidelines for ABGs have been 
published, 7 the indications for this analysis remain 
somewhat unclear.* It is important to recognize that 
arterial P02 (PaCb) is primarily the result of lung 
function and may bear no relationship to oxygen 
delivery to tissues or to tissue oxygenation. The 
numerous problems associated with the measure- 
ment and use of ABGs (Table 1 ) are generally well 
known. 9 " 27 

Table 1. Problems Associated with Arterial Blood Gas Analy- 

Pre-analytical errors (eg, air contamination, heparin dilution. 

Infection risk to patient (particularly with arterial catheter) 
Infection risk to clinician (particularly during performance of 

arterial puncture) 
Thrombosis and distal embolization (particularly with arterial 

Blood loss (particularly with arterial catheter) 
Intermittent nature of information 
Regulation: CLIA-88 

Laboratory evaluation accounts for about 25% 
of total critical care costs, and ABGs are the most 
frequently ordered laboratory test in critical care 
units. In a 1-year study (1987-1988) of the surgical 
intensive care unit (SICU) at the University of 
North Carolina, Chapel Hill. ABGs were ordered at 
a rate of 4.8/patient/day. :s The presence of an ar- 
terial line was the single most powerful predictor of 
the number of blood gases drawn per patient. This 
occurred regardless of the values for P a o:. PaCO:. 
APACHE II score, use of ventilators, or use of 
pulse oximeters (Fig. 2). 

Ventilator No Ventilator 


No Pulse 

Fig. 2. The average number of arterial blood gases per 
patient day for various patient types. ■ = no arterial line; 
■i = arterial line. (From data in Reference 29.) 

The presence of an arterial line also affects the 
transfusion requirements resulting from phlebot- 
omy for diagnostic tests in adults. Although the con- 
sequences of excessive phlebotomy for diagnostic 
testing are well appreciated in neonates, phlebotomy 
blood loss can also be a problem in adults. In a 
study of 100 patients at Beth Israel Hospital in Bos- 
ton, the presence of an arterial line was an im- 
portant predictor of blood loss from phlebotomy 
(Fig. 3). 29 Critically ill patients with an arterial line 
had blood drawn at a rate twice that of patients who 
did not have a line. Nine of the 50 critically ill pa- 
tients who had an arterial line present at some point 
in their hospitalization had more than 1.000 mL of 
blood drawn. Thirty-six patients were given trans- 
fusions at some point in their hospital course, and 
24 (67%) of these were critically ill patients with 
arterial lines. 

ICU with 

ICU without 



Fig 3. Total volume of blood loss from phlebotomy ac- 
cording to patient type, with and without arterial line. 
(From data in Reference 29.) 

Several things might be done to decrease the 
number of ABG analyses (and other laboratory 
testing) in critically ill patients. First, it might be 
possible to substitute clinically accurate and precise 
invasive and noninvasive technology for some di- 
agnostic tests. Second, multiple diagnostic tests 
could be performed on a single sample of blood. Is 
For example, some commercially available analyz- 
ers suitable for critical care require only a single 
sample for measurement of blood gases, pH, elec- 
trolytes, hematocrit, and blood sugar. Third, and 
perhaps most important, strict guidelines can be 
used to improve the appropriateness of diagnostic 
testing. This is illustrated in a study by Beasley et 
al conducted in the SICU at the University of Vir- 
ginia Health Sciences Center. 30 Procedural, clin- 




ical, and therapeutic indicators for ABG analysis 
were established. A forced interaction between res- 
piratory therapists and nurses was established by 
having therapists responsible for ABG analysis. 
This interaction between therapists and nurses re- 
garding ABG analysis has resulted in a significant 
improvement in the appropriateness of ABG analy- 
sis obtained by the nurses (Fig. 4). 


1 month 

1 year 

Fig. 4. Results of a forced interaction between res- 
piratory therapists and nurses ^ on the percentage 
of appropriate blood gas requests by nurses; note that 
the percentage of appropriate blood gas requests by 
nurses increased over time following the imple- 
mentation of therapists' measuring ABG. (From data in 
Reference 30.) 

One frequently unappreciated aspect of inter- 
mittent ABG analysis is spontaneous variability (ie, 
changes in ABGs without any intervention or inter- 
action with the patient). In work done in the SICU 
at York Hospital, we found a relatively high co- 
efficient of variation (CV) in P u o: among 4 ABGs 
drawn at 20-min intervals in stable mechanically 
ventilated trauma patients who were not disturbed 
during the study period (%CV of 5.1 ± 3.2%, me- 
dian 3.6%, 95th percentile 9.8%). 31 Similar findings 
have been reported by Thorson et al. 32 The clinical 
importance of these results is illustrated in Figure 
5. Changes in P a o: within the limits of spontaneous 
variability and without an associated clinical 
change can be the result of spontaneous variability. 

The gold standard for measurement of hemo- 
globin oxygen saturation (0 2 Hb) is CO-oximetry. 33 
CO-oximetry uses multiwavelength spectrophotom- 
etry to measure 2 Hb, deoxyhemoglobin (HHb). 
carboxyhemoglobin (COHb). and methemoglobin 
(metHb). Fractional oxyhemoglobin (F0 2 Hb) is 
then calculated as 

40 60 80 100 120 140 160 
Measured PaC>2 (torr) 

Fig. 5. Limits of spontaneous variability of P a o2 in 
stable, mechanically ventilated trauma patients who 
were not disturbed during the evaluation period. (From 
Reference 31, with permission.) 

An alternative method is to calculate hemoglobin 
oxygen saturation (functional saturation): 


: Hb 
OHb + HHb 

FO,Hb : 


Q,Hb + HHb + metHb + COHb 

This method is discouraged because it overlooks 
the effects of dyshemoglobins and thus overes- 
timates the true oxygen saturation. S0 2 can also be 
calculated from the measured P02 using an empir- 
ical equation for the oxyhemoglobin dissociation 
curve; this is commonly done automatically by 
software in blood gas analyzers. Clinically impor- 
tant errors can result from use of a calculated S0 2 
in applications such as the shunt equation. 34 " 37 It 
should never be assumed that F0 2 . S0 2 . and calcu- 
lated S0 2 are equivalent. 

There are several errors related to CO-oximetry. 
Due to differences in the light absorption spectra of 
fetal hemoglobin (HbF) and adult hemoglobins, the 
presence of HbF may result in false elevations of 
COHb. 3S Correction factors for the presence of HbF 
have been published, 30 but these may not be ap- 
propriate under all conditions (such as trans- 
fusion). 4 " It has also been shown that CO-oximetry 
yields falsely elevated metHb levels in the presence 
of lipid infusions. 41 

Point-of-Care Testing 

Interest has increased recently in rapid bedside 
assays of blood gases, pH, electrolytes, and hema- 




tocrit (or hemoglobin). 42 " 4 ' 1 This testing is often re- 
ferred to as point-of-care (POC) testing. Perhaps 
the most common example is that of bedside glu- 
cose monitoring. 46 The GEM-6/GEM-Stat (Mal- 
linckrodt Sensor Systems) is an analytical system 
that uses miniaturized electrodes to measure blood 
gases and electrolytes and electrical thermocon- 
ductivity to estimate hematocrit. 47 A disposable ana- 
lytical pack can process 50 samples over an active 
service life of up to 8 hours (and up to 36 hours 
when a standby mode is used). The instrument can 
manually analyze discrete blood samples and can 
aspirate blood from a cardiopulmonary bypass ma- 
chine. The system performs an initial 2-point cal- 
ibration that is repeated each hour and a 1 -point 
calibration with each measurement. Quality control 
solutions are provided by the manufacturer. The re- 
quired sample volume is 2 mL, and processing time 
is 130 seconds. 

The GEM-6 system was evaluated by Bashein et 
al 47 in 9 patients during open heart surgery. They 
found that the hematocrit and ionized calcium 
measurements were sufficiently accurate to replace 
conventional laboratory measurements. Although 
they found relatively wide limits of agreement for 
Po : (-4.2 to 24.6 torr), they found that the limits of 
agreement were better for venous samples (-8 to 4 
torr). This suggested that the instrument performed 
better at the lower end of the range for Po:- These 
limits of agreement for Po : may be too wide for use 
in critically ill patients. One major concern with 
this system is that it does not allow continuous 
monitoring (as do pulse oximetry or invasive Po: 

The future of POC testing in critical care is un- 
certain. Issues of quality control and proficiency 
testing are unresolved. If these devices are less ac- 
curate and precise than conventional laboratory 
measurements, the results may be of limited clin- 
ical usefulness. Although manufacturers of these 
devices advertise their simplicity, their perform- 
ance in the hands of persons without laboratory 
training (eg, persons without training in respiratory 
care or laboratory technology) may be subopti- 
ma j 48,49 p urmer> tne j ssues f cost-effectiveness 
have not been adequately addressed. The principal 
motivation for use of POC devices is to provide 
rapid accurate results at low cost. Whether this is 
possible in critical care remains to be seen. 

Continuous In-Vivo Evaluation of 
Arterial Oxygenation 

There is currently much interest in the critical 
care community in a system to continuously mon- 
itor arterial blood gases and pH. 50 " 53 Several man- 
ufacturers now have FDA approval for such sys- 
tems. To measure Po:, these systems use either a 
Clark electrode or an optode. In-vivo electrode sys- 
tems to continuously measure P02 have been de- 
scribed. However, in general these systems have 
been abandoned for several reasons (Table 2). 54 " 59 

Table 2. Disadvantages of a Clark Electrode System for Con- 
tinuous In-Vivo Monitoring of Oxygenation 

The size of the electrode requires an 18-gauge or larger arterial 

Drift results in systematic errors. 

The patient cannot be adequately isolated from the system. 
The electrode consumes oxygen. 
Accuracy is affected by blood flow. 
The electrode is thrombogenic. 
Accuracy is affected by deposition of blood constituents on the 


Current generation in-vivo blood gas systems 
use fluorescent optodes, which are optical bio- 
sensors first developed by Lubbers and Opitz, who 
coined the word optode (Fig. 6). The physical and 
chemical characteristics of these systems have been 
described in detail elsewhere. 60 " 63 Although these 
systems are only now being introduced into critical 
care medicine, they are in common use during 
cardiopulmonary bypass and extracorporeal mem- 
brane oxygenation (ECMO). The optode consists 
of a miniaturized probe containing a fluorescent 
dye (Fig. 6.) The dye is capable of absorbing light 
of a specific wavelength and rapidly re-emitting the 
light at a longer wavelength (lower energy). Some 
molecules and ions (eg, 2 . CO : , H + ) are capable of 
accepting energy from the fluorescent dye. This pro- 
cess, known as quenching, decreases the amount of 
emitted energy relative to the concentration of : 
(or other specific molecule) that is present. Photo- 
sensors are used to quantify the amount of quench- 
ing, and a microprocessor is used to translate the sig- 
nal into a display of Po 2 - Optode systems are pre- 
sently available to measure Po;, Pco:- and pH. In the 
future, additional optodes may become available to 
measure other constituents of whole blood in vivo. 






te Permeable 

? 1 

H Indir.uor Dye | 

Emitted Light 


Fig. 6. A schematic illustration of an optode to measure 
Po2- The optode is miniaturized so that it can fit through 
an arterial catheter. See text for details. 

The principal advantages of optode systems 
(compared to polarographic systems) are that they 
can be electrically isolated from the patient and can 
be miniaturized to fit through a 20-gauge arterial 
catheter without affecting blood pressure measure- 
ments or other catheter functions. Minimal re- 
quirements for a blood gas monitor in conjunction 
with an arterial catheter have been suggested by 
Shapiro (Table 3). 50 Several approaches to blood 
gas monitoring can be taken in conjunction with an 
arterial catheter. The first approach uses a probe 
that passes through the arterial catheter and resides 
directly within the arterial lumen. With the second 
approach, the optode system is connected to the 
proximal arterial line but does not pass through the 
catheter. With this method, when blood gas and pH 
values are desired, blood is drawn into a chamber 
containing the optodes. After analysis, the blood is 
flushed back into the artery. Thus frequent (but not 
continuous) blood gas measurements are possible 
without blood loss. 

Several manufacturers have subjected optode 
systems to rigorous clinical trials for purposes of 
gaining FDA approval before commercial release 
of the technology. However, only a few studies 
evaluating these devices have appeared in the peer- 
reviewed literature. 64 " 67 Shapiro et al evaluated an 
intra-arterial optode blood gas system in dogs and 
humans. 64 Although the system was used for as 
long as 25.5 hours in 3 patients, monitoring was 
terminated in 5/12 (42%) due to system fault. Ac- 
ceptable bias and precision were found for pH and 
Pco:- Although bias was acceptable for P02 meas- 
urements, the precision was marginal at best (Fig. 
7). Similar results have been reported by others. 65 

Table 3. Minimal Requirements for Blood Gas Systems Used 
in Conjunction with Arterial Blood Catheters* 

Must accurately measure P02, Pco:, pH, and temperature, with 
a rapid response time 

Must not have any interference from substances normally 
found in arterial blood 

Must operate with a 20-gauge arterial catheter without af- 
fecting continuous blood pressure measurement, obtaining 
blood samples, or any other functions of the arterial cathe- 
ter system 

Must be biocompatible and nonthrombogenic 

Must be simple to operate and maintain 

Must be able to withstand the rigors and abuse of everyday 
practice within the ICU 

Must be accurate and precise for at least 72 hours 

Must not be adversely affected by reductions in local blood 
flow or perfusion 

Must not be adversely affected by changes in hemodynamics 

Must be cost-effective 

*Adapted from Reference 50, with permission. 

Several considerations are important when eval- 
uating the accuracy of these devices. Although the 
gold-standard comparison is results of ABG analy- 
sis, laboratoiy accuracy should be distinguished 
from clinically useful accuracy. It is conceivable 
that a device that does not meet laboratory ac- 
curacy standards might still be a useful clinical 
evaluation device in the ICU. Although in-vivo 
blood gas monitoring devices may (or may not) 
sacrifice some of the accuracy of the blood gas an- 
alyzer, in-vivo devices will not be subjected to the 
preanalytical errors that occur with ABGs. 

Despite considerable clinical interest, the future 
of in-vivo measurements is unclear. One problem 
with these systems is their cost. Although blood 
gas analyzers may be equally expensive, blood gas 
analyzers are not dedicated to a single patient. Al- 
though manufacturers are striving to design user- 
friendly systems, the amount of technical attention 
that these systems are likely to require in a busy 
ICU remains to be determined. The life of the op- 
tode and its effect on the function of the arterial 
catheter system (eg, blood pressure measurements) 
in a busy ICU also remain to be seen. The quality 
control and quality assurance requirements for clin- 
ical and regulatory purposes are currently unclear. 
It is likely that there will also be interest in using 
these systems to achieve closed-loop ventilation. To 
avoid the mistakes that have been made with pulse 




Average P0 2 from ABG & IBGS 

ABG PQ 2 (torr) 

Fig. 7. A. Relationship between P a c>2 and P02 from an intra-arterial blood gas system (IBGS) in dogs. B. Bias and precision 
of IBGS compared with P a o2 in a human study; note relatively wide limits of agreement. (From data in Reference 64.) 

oximetry, the effect of these systems on outcome 
needs to be evaluated. Whether the benefits of con- 
tinuous blood gas and pH measurements will out- 
weigh the costs and technical support required re- 
mains to be seen. 

Pulse Oximetry 

Pulse oximetry, a technology unavailable until 
the mid-1980s, is probably now common in every 
critical care unit of the modern world. More than 
35 companies manufacture pulse oximeters; 1989 
worldwide sales were estimated at over 65,000 units 
(> $200 million). 68 Probably no other monitoring 
technology has so quickly and completely infil- 
trated and permeated the practice of critical care. 
With virtually no scientific evaluation of its impact 
on outcome, continuous pulse oximetry appears to 
have become a standard of care for critically ill pa- 
tients. Although this practice may be debatable, the 
reality is that pulse oximetry has become an inte- 
gral component of many critical care monitoring 
systems. Pulse oximeters are available as small hand- 
held units, as semiportable units with alarms and 
waveform display, in combination with other mon- 
itors such as capnography and transcutaneous Pccb. 
incorporated into mechanical ventilator systems, 
and as part of the bedside critical care monitoring 
system. Pulse oximetry is used in prehospital care, 69 
in the emergency department, 70 in neonatal and pe- 
diatric critical care, 7 ' in adult critical care, and in 
general patient care wards. 72 

The technical and physiologic aspects of pulse 
oximetry have been reviewed in detail else- 
where. 6873 " 77 Two wavelengths of light (660 nm and 

940 nm. Fig. 8) are passed through a pulsating vas- 
cular bed. This is accomplished by using two light- 
emitting diodes (LEDs) and a photodetector. There 
is some error in the wavelength of light emitted by 
the LEDs (± 30 nm) that can affect accuracy. Also, 
the photodetector is not specific (ie, it will respond 
to any wavelength of light, which can result in in- 
terference). Although some of the light emitted from 
the LEDs is absorbed by each constituent of the tis- 
sue, the only variable absorption is due to arterial 
pulsations. This is translated into a plethysmo- 
graphic waveform. The ratio of the amplitudes of 
these two plethysmographic waveforms is trans- 
lated into a display of oxygen saturation. A probe is 
used to pass light from the LEDs through a pul- 
sating vascular bed. A variety of probes is available 
and includes finger, ear, nasal, and foot probes, in 
disposable and reusable designs. Although most 
pulse oximeters use transmission oximetry (ie, the 
light from the LEDs is transmitted through the tis- 
sue, and the photodetector is opposite the LEDs), 
other designs use reflectance oximetry 78 " 80 (ie, the 
light from the LEDs is reflected from the tissue, 
and the photodetector is on the same side of the tis- 
sue as the LEDs). 


A number of limitations of pulse oximetry 
should be recognized, appreciated, and understood 
by everyone who uses pulse oximetry data. 8 ' 83 
Most pulse oximeter errors result from too little 
signal (eg, low perfusion, improper probe place- 
ment) or too much noise (eg, motion, ambient 
light). 68 









Wavelength (nm) 

Fig. 8. Light absorption spectra tor Hb and : Hb. The 
wavelengths of light indicated are those used by pulse 

Accuracy — Pulse oximeters use empirical calibra- 
tion curves developed from studies of healthy vol- 
unteers. Many clinical evaluations of the accuracy 
of pulse oximeters have been published, each com- 
paring the pulse oximeter saturation (S P 02) to the 
saturation of a simultaneously obtained arterial 
sample measured on a CO-oximeter. At saturations 
> 80%, the accuracy of pulse oximetry is about ± 4- 
5%. Below 80%, the accuracy is less, but the clin- 
ical importance of this lesser accuracy is ques- 
tionable. To appreciate the implications of the limits 
of accuracy of pulse oximetry, one must consider 
the oxyhemoglobin dissociation curve (Fig. 9). If 
the pulse oximeter displays a S p c>2 of 95%, the true 
saturation could be as low as 90% or as high as 
100%. If the true saturation is 90%, the Po : will be 
about 60 torr. However, if the true saturation is 
100%, one does not know how high the Po 2 might 
be. Clinically, pulse oximeters should be considered 
"desaturation meters," with a low S P 02 considered to 
reflect a low S p o: until proven otherwise. 68 

Differences among devices and probes — The 

pulse oximeter is unique as a respiratory monitor in 
that it requires no user calibration. However, man- 
ufacturer-derived calibration curves programmed 
into the software vary from manufacturer to man- 
ufacturer and can vary among pulse oximeters of a 

given manufacturer. As pointed out above, the LED 
output can vary from probe to probe. The result of 
these factors is that the accuracy of pulse oximetry 
varies among devices. 84 " 90 For these reasons, the 
same pulse oximeter and probe ideally should be 
used for all S p o: determinations on a given patient. 

P0 2 (torr) 

Fig. 9. Oxyhemoglobin dissociation curve. If a pulse ox- 
imeter indicates an S P 02 of 95%, the true : Hb may be 
as low as 90% or as high as 100%, reflecting a wide 
range of P a 02 values. 

Penumbra effect — If the finger pulse-oximeter 
probe does not fit correctly, light can be shunted 
from the LEDs directly to the photodetector. 9 ' 92 
Theoretically, this will cause a falsely low S P 02 if 
saturation is greater than 85% and a falsely elevat- 
ed S p o: if saturation is less than 85%. 

Dyshemoglobinemias — Because commercially avail- 
able pulse oximeters use only two wavelengths of 
light, they are able to evaluate only ; Hb and 
HHb. COHb and metHb concentrations are as- 
sumed to be low. Abnormal elevations of COHb 
and metHb both result in significant inaccuracy in 
pulse oximetry, 93 " 96 and pulse oximetry should not 
be used when elevated levels of these are present. 
Although fetal hemoglobin may affect the accuracy 
of CO-oximetry, it does not seem to affect the ac- 
curacy of pulse oximetry. 38,97 

Endogenous and exogenous dyes and pigments — 

Vascular dyes can affect the accuracy of pulse ox- 
imetry, with methylene blue having the greatest ef- 
fect. 98 ' 99 Nail polish can also affect the accuracy of 
pulse oximetry and should be removed before pulse 
oximetry is begun. 100 Interestingly, hyperbilirubi- 




nernia does not appear to affect the accuracy of 
pulse oximetry."" 102 

Skin pigmentation — Several studies have found 
that accuracy and performance of pulse oximeters 
are reduced by deeply pigmented skin." 11 " 1 " 5 

Perfusion — To function correctly, pulse oximeters 
require a pulsating vascular bed. Under conditions 
of low flow (eg, cardiac arrest or severe peripheral 
vasoconstriction), pulse oximetry becomes unreli- 
able. Under these conditions, an ear probe may be 
more reliable than a finger probe. 68 

Anemia — Although pulse oximeters are generally 
reliable over a wide range of hemoglobin levels, 
they become less accurate and reliable in condi- 
tions of severe anemia (Hb < 8 g/dL at low satura- 
tions and hematocrit < 10% at all saturations). 106,107 

Motion — Motion of the probe can produce consid- 
erable artifact and unreliable and inaccurate pulse 
oximetry readings. 68 This problem can sometimes 
be corrected by using an alternate probe site (such 
as the ear or toe rather than the finger). 

High intensity ambient light — Because the photo- 
detector of the pulse oximeter is nonspecific, high- 
intensity ambient light can produce interference. 
The problem can be corrected by wrapping the 
probe with a light barrier. 68 

Abnormal pulses — Venous pulses and a large di- 
crotic notch have been shown to affect the accuracy 
of pulse oximetry. 108 " 110 

Safety and Usefulness 

Pulse oximeters are generally considered safe. 
However, bums from defective probes and pressure 
necrosis have been reported. 1 """ 4 

It is our clinical impression that routine continu- 
ous pulse oximetry during mechanical ventilation 
detects few clinically important desaturation events 
that are not otherwise detectable. Although only a 
weak case can be made for routine use of pulse ox- 
imetry, its use should not be withheld when it is in- 
dicated. Pulse oximetry is indicated in unstable pa- 
tients likely to desaturate, in patients receiving 
therapeutic interventions that are likely to produce 
hypoxemia (such as bronchoscopy), and in patients 

having interventions likely to produce changes in 
arterial oxygenation (such as changes in F102 or 
PEEP). 2 The pulse oximeter is probably no better at 
detection of a disconnect than the alarms already 
available on the ventilator. The pulse oximeter may 
actually be more likely to produce annoying false- 
positive alarms.' 68 and there may be a relatively 
long period between disconnection and de- 
saturation (particularly if the P a c>2 is high before the 

Although pulse oximetry may improve the 
detection of desaturation, no study demonstrating 
that pulse oximetry makes a difference in morbid- 
ity and mortality has yet been published. 68 Unfor- 
tunately, the acceptable level of desaturation is not 
known. The potential scenario described by 
Keats" 5 is to be avoided: "Without (outcome data 
on pulse oximetry) I can envision a subpopulation 
of this country walking around without their front 
teeth because of urgent intubation when an ox- 
imeter read less than 90%." Keats further observes 
"cases that only two years ago would have been 
classified as inadequate ventilation . . . now . . . 
have oximeters in place and they show no desatura- 
tion." 115 

Jubran and Tobin evaluated the use of pulse ox- 
imetry in titrating supplemental oxygen in 54 crit- 
ically ill ventilator-dependent patients." 14 In white 
patients, they found that an S P 02 of 92% was re- 
liable in predicting a P a o: ^ 60 torr (Fig. 10). How- 
ever, in black patients, an S p o: of 95% was re- 
quired. Although this method is useful for titrating 
to a level of arterial oxygenation that does not pro- 
duce hypoxemia, it does not eliminate the need for 
periodic ABG. When pulse oximetry is used to ti- 
trate F102, the final F102 setting should always be 
confirmed by an ABG. Interestingly, no similar 
study has been done to evaluate the reliability of 
pulse oximetry for titrating PEEP. 

If pulse oximetry is to be clinically useful, it 
must have a low failure rate. Intraoperative pulse 
oximetry failure was evaluated by Freund et al." 6 
Overall, they found a failure rate of < 5%. Pulse 
oximetry failures tended to be greater in older and 
sicker patients and during longer surgical pro- 
cedures. To our knowledge, pulse oximetry failure 
in the critical care unit has not been studied. We 
suspect that the incidence of pulse oximetry fail- 
ures is much higher in the ICU than in the rel- 
atively controlled environment of the operating 








o 70- 

D 'V* 







30 _ 

Spo 2 94% 



90 -I 

_ 80 

1 70^ 

S 60 









5 70- 






S p o 2 92% 


. (198) 
. (107) 

_ 80- 
g 70- 
m fin 

• • 


• •• 


Q? DU 





S p o 2 90% 

S p o 2 95% 

Fig. 10. P a o2 values at 4 target S P 02 levels. The open circles represent white patients, and the closed circles represent 
black patients. In white patients, a target S P 02 of 92% satisfactorily predicted a P a 02 of 60 torr, whereas a target S p c>2 of 
95% was required to predict a P a 02 of 60 torr in black patients. (From Reference 104, with permission.) 

room. Perhaps the most frequent cause of pulse ox- 
imetry failure in the ICU is accidental dis- 
connection of the probe from the patient. 

The plethysmographic waveform provided by 
the pulse oximeter has been described as useful for 
evaluating collateral blood flow to the hand and 
systolic blood pressure.' 17 " 121 However, the use- 
fulness of this application is currently unknown. 

Transcutaneous P02 

Transcutaneous P02 (PtcCh) became clinically 
available in the mid- 1 970s. |:: Although this tech- 
nology is used in the neonatal ICU, it has limited 
use in the care of adult patients (Table 4). Since the 
introduction of pulse oximetry, P tC 02 monitoring 
has also declined in the neonatal ICU. Most com- 
mercially available transcutaneous monitors use a 
combination PtcCh/PtcCO: electrode. 

Table 4. Limitations of P| C o Momtcnn^ 

Frequent calibration is required. 

Frequent changes of electrode position are required. 

Relatively long equilibration time is required following elec- 
trode placement. 

Insufficient electrode temperature may adversely affect per- 

Performance may be suboptimal over poorly perfused areas. 

PtcO: tends to underestimate P a O;- 

Compromised hemodynamic status causes an underestimation 

of PaOz- 

Heated electrode may cause skin to blister. 
PtcO: may underestimate P a o: during hyperoxemia. 
Frequent membrane/electrolyte changes and electrode main- 
tenance are required. 
Performance is more reliable in neonates than in adults. 

*Adapted from Reference 2. with permission. 




The PtcOz electrode uses a polarographic princi- 
ple similar to that used in blood gas analyzers. 122 124 
To produce a P t cO: approximating P a 02- the elec- 
trode must be heated to approximately 44°C. The 
close relationship between P a o: and Pro: in neo- 
nates is the result of a complex set of physiologic 
events. Simply stated, the increase in P02 caused by 
heating roughly balances the decrease in Po: caused 
by skin oxygen consumption and the diffusion of 
oxygen across the skin. The close relationship be- 
tween P a o2 and PtcCb that occurs in neonates is 
probably more coincidental than physiologic. Fail- 
ure to recognize this creates the illusion that PtcO: is 
the same as P a 02- 

In adults, the P tC 02 is frequently less than P a 02. 
and the relationship between Pu.02 and P a 02 in 
adults is too variable to allow P tC 02 to be of much 
use for tracking P a o2- This has been demonstrated 
in several recent studies, 125 ' 127 including one large 
multicenter study (Fig. II). 127 

120 180 

Plc0 2 (lorr) 

Fig. 11. The relationship between Ptc02 and P a o2 in a 
multi-institutional study with 723 comparisons; note that 
Ptc02 grossly underestimates P a o2 when P02 > 60 torr. 
(Adapted from Reference 127, with permission.) 

The relationship between P tC 02 and P U 02 is given 
by the P tC 02 index (PtcCb/PaCb)- 124 This index de- 
creases with age due to progressive changes in skin 
thickness and perfusion (Table 5). 124 The P1CO2 in- 
dex also decreases with decreases in tissue oxyge- 
nation. P tC 02 is affected by perfusion and may re- 
flect oxygen delivery (the product of cardiac output 
and arterial oxygen content) to the skin under the 

electrode. 11 In other words, P tC 02 tracks P a o2 when 
perfusion is adequate, and P tc 02 tracks cardiac out- 
put when P a o2 is adequate. In fact, Pt c 02 has been 
used in adults to monitor the results of vascular sur- 
gery, the intent being to evaluate perfusion rather 
than P02 per se.' 2s A P tC 02 index < 0.7 reflects in- 
adequate tissue oxygenation. 126 

Table 5. Changes in P lc o: Index (P[ C o:/PaO;) with Age* 

Age Group 

Pti.02 Indexf 

Older adult 






* Adapted from Reference 1 24, with permission. 

t An index of 1 .0 means that P tc o: = PaOs, an index < 1 .0 

means that P tc O: < PaOi- and an index > 1 .0 means that P tc o; 

> Pa02- 

Both canine 129 "' 3I and human 132 " 137 studies have 
shown that P tC 02 decreases with a decrease in car- 
diac output and peripheral perfusion. In dogs, P tC 02 
decreases with hemorrhage (Fig. 12) and increases 
when the shed blood is reinfused (Fig. 13). 129 " 131 
With shock and decreased peripheral perfusion, the 
P tC 02 index decreases, and the magnitude of the de- 
crease is determined by the magnitude of the de- 
crease in perfusion. 132 " 135 

150 300 450 600 
Hemorrhage Volume (ml_) 

Fig. 12. Changes in transcutaneous P02 (• — ). mixed 
venous P02 (o ----). and cardiac output (± — ) during ac- 
tive hemorrhage in dogs. | = standard deviation. (From 
Reference 129, with permission.) 




150 300 450 600 750 

Reinfusion Volume (ml_) 

Fig. 13. Changes in transcutaneous P02 (• — ), mixed 
venous P02 (o --), and cardiac output (a — ) during 
tluid infusion in hypovolemic dogs, i = standard De- 
viation. (From Reference 129, with permission.) 

The use of P tC 02 monitoring during cardiac arrest 
and resuscitation has been reported. | - ,| -'- ,:! - l - ,4 - l - ,: > Dur- 
ing CPR in 5 patients, Tremper and Shoemaker' 32 
reported a P tc O; of 0-3 torr, a mean (SD) cardiac in- 
dex of 0.8 (0.2) L • miir 1 ■ nr : , a P a o 2 of 40 ( 12) torr, 
and an oxygen delivery of 59 (41) L- min" 1 ■ m" : . 
Abraham et al 134 - 136 - 137 reported a P to o 2 of 0-40 torr 
during CPR (Fig. 14). 

CPR slopped 
CVP measurement 
Pa0 2 206 | CPR restarted 
i CPR restarted |r 

* CPR stopped 


Fig. 14. Changes in transcutaneous and conjunctival 
P02 during resuscitation. (From Reference 137, with 

Although reports in the early 1980s promoted the 
use of PtcO: as an indicator of the effectiveness of 
CPR, this has not become common practice for sev- 

eral reasons. First, PtcO: monitoring is technically 
difficult — particularly during CPR. Second, an equil- 
ibration period of 10-15 minutes is required after 
placement of the P tc o: electrode, which limits its use 
early in the resuscitation unless the electrode was in 
place prior to the cardiac arrest (which is unlikely). 

Mixed Venous Blood Gas Assessment 

Blood obtained from the central circulation is 
used to assess mixed venous oxygenation. Al- 
though either P v O: or Sv02 can be used, mixed ve- 
nous oxygenation has conventionally been assessed 
for SvOz- The mixed venous oxygenation has been 
commonly thought to represent tissue oxygena- 
tion. 138139 PvO: < 28 torr has been shown to be as- 
sociated with a high likelihood of hyperlacticemia 
and death. 14 " 

Intermittent sampling of mixed venous blood is 
associated with the same preanalytical errors as ar- 
terial blood sampling. An additional error is related 
to withdrawal of blood from pulmonary artery cath- 
eters. If the catheter is positioned in a distal branch 
of the pulmonary artery, rapid aspiration can con- 
taminate the mixed venous blood with pulmonary 
capillary blood (Fig. 15). I39141 Although the im- 
portance of this has been questioned, 142 it seems 
prudent to use a slow aspiration rate (< 1 mL/30s) 
when blood is sampled from a pulmonary artery 
catheter. 134 

Pulmonary Artery 

Pulmonary Artery 
(mixed venous blood) 


Pulmonary Capillary 

Fig. 15. Schematic illustration of the pulmonary artery 
catheter in relation to the pulmonary capillary bed. 
Rapid withdrawal of blood from the distal port of the pul- 
monary artery catheter can result in contamination of 
the sample with arterial blood. 

Several points are important related to the ability 
of mixed venous oxygenation to reflect tissue oxy- 
genation. SvO: depends on the oxygen content de- 
rived from many vascular beds and thus does not 
necessarily reflect well the S v o: of any individual 




vascular bed. 138,139,143 With septic shock, peripheral 
arteriovenous shunts open, resulting in an SvO: 
higher than expected. In some patients, tissue oxy- 
gen consumption may decrease in proportion to a 
decrease in oxygen delivery; in such cases, Sv02 
may not be an adequate indicator of tissue oxy- 
genation. SvO: is also falsely elevated (ie, does not 
reflect tissue oxygenation) in patients with ven- 
tricular septal defect or cyanide poisoning. SvO: can 
be mathematically derived from the Fick Equa- 
tion:' 34144 




where Vo? is oxygen consumption and Do: is oxygen 

The relationship Vo : /Do: is known as the oxygen 
utilization coefficient or the oxygen extraction ra- 
tio. SvO: is tnu s determined by the balance between 
Vo : and Do: (Fig. 16). 144 


.2 60 


^ 50 


o .„ 

= 40 



= 30 

:t ,. 


■ "•' • 

<S 20 





i i i i 



50 60 70 80 90 100 
Oximetry Sv 02 

Fig. 16. The relationship between mixed venous oxy- 
gen saturation and the oxygen utilization ratio, r = 
p < 0.001. (From Reference 144, with permission. 


True mixed venous blood is obtained from the 
pulmonary artery. Normally, oxygen saturation is 
slightly higher in the inferior vena cava than in the 
superior vena cava due to the relatively low levels 
of oxygen extraction relative to perfusion in the re- 
nal and mesenteric vasculature. I34 However, during 
periods of stress (eg. shock, hypovolemia, exercise), 
oxygen saturation in the inferior vena cava drops, so 
that saturation in the superior vena cava is greater 
than that in the inferior vena cava (Fia. 17). This oc- 

curs as the result of decreased renal and splanchnic 
perfusion during these conditions. Thus, central ve- 
nous blood other than that obtained from the pul- 
monary artery may inadequately reflect true mixed 
venous blood. Although mixed venous blood sam- 
ples ideally should be obtained from the pulmonary 
artery, a high correlation between pulmonary shunt 
calculations using pulmonary artery samples and 
central venous samples (eg, right atrial or superior 
vena caval) has been reported (Fig. 18).' 4? 



Fig. 17. Hypothetical changes in venous oxygen sat- 
uration of blood in the superior vena cava (SVC), in- 
ferior vena cava (IVC), and pulmonary artery (PA) un- 
der normal conditions and conditions of shock. (Con- 
cept adapted from Reference 139.) 

Venous Oximetry 

The ability to continuously measure mixed ve- 
nous oxygen saturation via oximetric methodology 
became commercially available in the early 
1980s. 146 This system uses a microprocessor, an op- 
tical module with light sources and photodetectors. 
and a flow-directed pulmonary artery catheter. 146 " 148 
Using fiberoptics, wavelengths of light between 
650 nm and 1,000 nm are pulsed into the pul- 
monary artery (Fig. 19). This light is then reflected 
from red blood cells in the pulmonary artery and 
returned to the optical module via another fib- 
eroptic bundle. SvO: is then determined by the ratio 
of transmitted and reflected light. Before insertion, 
the system is calibrated using an in-vitro calibra- 
tion standard, and calibration can be updated pe- 
riodically by in-vivo calibration using a CO- 
oximetry-determined SvO:- Factors that affect the 
measurement of SvO: using this method include 
temperature. pH. blood flow velocity, hematocrit. 






- Qsp/Qt 

y = 0.838x + 0.020 / / 

/ • / 


r = 0.960 / / 
n = 55 /* / / 

J V / 


/ *9* / 


i i i i i 

1 1 1 







AQ Rn /Q 

Fig. 18. The relationship between pulmonary shunt values calculated using pulmonary artery and central venous blood 
samples; left figure shows the regression line, with 95% confidence limits of solitary samples from separate patients; right 
figure shows the relationship between simultaneous changes in shunt values obtained from the central venous (CV) and 
pulmonary artery (PA). (From Reference 145, with permission.) 

and occlusion of the catheter tip (eg, clot or vessel 
wall). 146 " 149 Although CO-oximetry is affected by 
lipid emulsion infusions, such infusions do not ap- 
pear to affect in- vivo venous oximetry. 150 

Pulmonary Artery 

To Oximeter 

Receiving Fiberoptic 

Transmitter Fiberoptic 

Distal Lumen 
Fig. 19. Diagram of mixed venous oximetry catheter. 

Three systems are commercially available to 
measure SvOt by oximetry. Each uses a different 
method to deal with interference and drift. 151 

Edwards Sat-One Catheter 

(Edwards Critical Care Division, Baxter Health Care 

Corp, Irvine CA) 

This system uses two reference wavelengths and 
one detecting fiberoptic filament. The system al- 
lows the user to periodically update the hematocrit 
and, in that way, control the effects of hematocrit 
on SvO: measurements. 

Oximetrix Opticath Catheter 
(Abbott Critical Care Systems, Hospital 
Products Division, Abbott Laboratories, 
North Chicago IL) 

This system uses three reference wavelengths and 
one detecting filament. The third wavelength sup- 
posedly improves the accuracy of the system in the 
face of physiologic changes such as in hematocrit. 

Spectramed Spectracath 

Catheter (Z1660-Spectramed, Oxnard CA) 

This system uses two reference wavelengths and 
two detecting filaments. In this case, the second de- 
tecting filament supposedly improves the accuracy 
of the system when hematocrit changes. 




Table 6. Studies Evaluating Correlation between Mixed Venous Oximetry and CO-Oximetry 

Nelson 144 
Fahey 148 
Rouby 151 

Baele 152 

Waller 1 " 
Gettinger l: 



Reinhart 157 

Hecker 158 

Schranz 159 

VanWoerkens 160 

Vaughn 1 " 1 

Rasanen 164 

Zaune 165 

39 adults 
86 adults 

1 1 adults 
8 adults 

12 adults 

16 adults 

13 adults 

10 dogs 
10 dogs 

12 adults 

15 adults 

38 adults 
37 adults 

32 adults 

33 adults 

16 infants 
6 adults 

46 adults 

10 adults 

15 adults 
15 adults 

). of Measurements 

























































Not provided 



Not provided 



Not provided 



A number of studies have evaluated the accuracy 
of SvO: catheter systems (Table 6). 144152 " 161 Most, 
but not all, 157 have found the 3-wavelength systems 
to correlate better with CO-oximeter S v o: than 2- 
wavelength systems. All of these papers report cor- 
relation coefficients, unfortunately, when accuracy 
and precision analysis (as described by Bland and 
Altman 162 ) might have been more appropriate. It 
has also been generally found that drift is less with 
3-wavelength systems. 151 " 154,155 " 158 However, if 
meticulous attention is paid to care of the catheter 
system (eg, in-vivo calibrations, updating the hem- 
atocrit setting), either system is probably accept- 
able. A recent study found that each of these three 
systems lacked precision, and that a change in S v o: 
of at least 7% must occur to be considered clin- 
ically important. 163 Unfortunately, it is frequently 

unappreciated that these systems require more tech- 
nical attention than other monitoring systems such 
as pulse oximetry or a conventional pulmonary ar- 
tery catheter. 

Regardless of their technical performance, the 
clinical usefulness of SvCh monitoring remains un- 
clear.' 61163 " 166 Although the accuracy and per- 
formance of these systems have been reported, rela- 
tively little has been reported regarding the effect 
on patient outcome. 

Dual Oximetry 

Dual oximetry is the simultaneous use of pulse 
and venous oximetry. 164 m Rasanen et a i 16416917 - 
have described the use of dual oximetry to de- 
termine the ventilation-perfusion index (VQI) I7: 
and oxygen extraction index (0 : EI). 169 




VQI is calculated using the following equation: 
[( 1.34)(Hb)( 1- S p o:) + ( 0.003 )(P A 0:>] 

VQI = 

[(1.34)(Hb)(l-SvO:) + (0.003 )(Pao : )1 

where Pao: is the alveolar Po: (assuming P a co: = 40 torr). 

The VQI assumes that pulmonary end-capillary 
blood is fully saturated with oxygen, and that the 
contribution of dissolved oxygen to oxygen content 
is relatively small. Further, VQI assumes that S P 02 ^ 
99%. VQI is supposedly an index of Qsp/Qt. 164 -' 71 ' 172 
2 EI is calculated using the following equa- 

(S p o: - Sv02> 

OEI = 

S p o: 

2 EI supposedly is an index of the oxygen extrac- 
tion ratio (oxygen utilization coefficient, Vo 2 / 
Do 2 ). 164 ' 169 Like the VQI. 2 EI assumes that the 
contributions of dissolved oxygen to total oxygen 
content are negligible. 

Several papers have evaluated the correlation be- 
tween VQI and Qsp/Qt 164 ' 171 and the correlation be- 
tween 2 EI and oxygen extraction ratio. I64J69 The 
correlation between 2 EI and oxygen extraction ra- 
tio has been found to be high (r > 0.9, Fig. 20). The 
correlation between VQI and Q sp / Qt, however, has 
been only modest (r = 0.77, Fig. 21). As seen in 
Figure 21, the scatter in the relationship between 
VQI and Q sp / Qt limits its usefulness. Although one 
study found VQI useful to determine optimal PEEP 

2 EI 

Fig. 20. Relationship between oxygen extraction index 
(0 2 EI) and oxygen utilization coefficient (0 2 UC )in 10 
patients with acute respiratory failure. (From Reference 
164, with permission.) 

in 14/17 patients, 170 to our knowledge this method 
has not become popular. Due to the many assump- 
tions made in the calculations of VQI and 2 EI. 
coupled with the relative inaccuracies of pulse ox- 
imetry and venous oximetry, VQI and 2 EI are 
probably at best crude indices of Q sp / Qt and oxygen 
extraction ratio. 

40 - 

35 - 


30 - 






O 2 °- 


15 - 





10 - 









Fig. 21. Relationship between Q sp /Qtand VQI in 10 pa- 
tients with acute respiratory failure. (From Reference 
164, with permission.) 

Oxygen Consumption 

One method of measuring Vo 2 is the reverse Fick 

Vo; =(Q)(CaO:-CvO:). 
where Q = cardiac output. 

This method has been shown to be useful in the de- 
termination of metabolic rate and caloric re- 
quirements in critically ill patients. 173174 There are 
several problems related to the use of this rela- 
tionship, particularly in critically ill patients in 
whom the intent is to evaluate the supply-depen- 
dence of oxygen utilization. 176 (Due to these prob- 
lems, it may be better to use indirect calorimetry to 
evaluate Vo 2 .) 


Accurate calculation of Vo 2 using the Fick Equa- 
tion requires accurate measurement of cardiac out- 
put (ie, Q), usually by thermodilution, Hb, 2 Hb of 
arterial and mixed venous blood, P a o 2 . and P v o 2 - 
Random errors and proportional errors in these 
measurements may result in significant errors in 
the calculated Vo. 




Pulmonary V02 

Vo 2 measured by the reversed Fick Equation is 
consistently less than that measured by indirect cal- 
orimetry. 177 " 17 '' This difference has been explained 
as the result of pulmonary oxygen consumption. 179 
Oxygen consumption within the lungs is usually 
low, but may be as much as 15% of the total body 
Vo 2 if the lung is infected. 

Mathematical Coupling of Data 

This issue relates to the statistical problems of 
using shared measurements to calculate Vo 2 and 
Do2 180-183 £ a02 and q are both used to ca i cu iate Vo 2 

and Do:- Measurement error in either of these var- 
iables will cause a related shift in both variables. 

Indirect Calorimetry 

The principles of indirect calorimetry have been 
described in detail elsewhere. 184185 The open circuit 
indirect calorimeter measures the concentrations 
and volumes of inspired and expired gases to de- 
termine Vo 2 (Fig. 22). Vo 2 is then calculated as 

r ( 1 - Fe02 - FEC02)(Fl02) -, 

V02 = [ Feo:J Vfe- 

( 1 - Feo 2 ) 

Several points are important for an open-circuit in- 
direct calorimeter to work properly: The F102 must 

be stable, the F102 must be < 0.60, the entire system 
must be leak-free, and the inspired and expired 
gases must be completely separated. With the 
closed-circuit indirect calorimeter, the patient re- 
breathes from a closed system, and the decrease in 
the volume of the system is Vo : (Fig. 23). Factors 
that affect the performance of the closed-circuit cal- 
orimeter include leaks, increased compressible vol- 
ume, and increased trigger effort during mechanical 
ventilation. A major advantage of the closed-circuit 
calorimeter is that it can be used with any Fio:- 



Fig. 23. Schematic diagram of a closed circuit indirect 
calorimeter. (From Reference 185, with permission.) 

Continuous Fick Cardiac Output 

The techniques of dual oximetry and indirect 
calorimetry can be combined to allow continuous 
estimation of cardiac output 18 ' 1 from the Fick Equa- 







Mixing Chamber 

Fig. 22. Schematic diagram of an open circuit indirect 
calorimeter. (From Reference 185, with permission.) 

(V 0: ) 

Q = - 


If the partial pressures of : are ignored. 1 " then 
Q = V 0: [(S p0 ;-SvO:KHb)(1.34)]. 

Failure to include the effects of P02 results in var- 
iable degrees of inaccuracy in this method. To min- 
imize this error, several approaches have been 
used. The P a 0: and PvO: can be back-extrapolated 
from the S p o 2 and SvO : .' ss l8g This method makes 
assumptions about the oxyhemoglobin dissociation 
curve and may be a grossly inaccurate estimate of 
Po 2 - Another method involves periodic measure- 
ment of P a o: and Pvo 2 and then assumption that 




these remain constant (again, often an incorrect as- 
sumption). 190 Perhaps the greatest inaccuracy re- 
lated to continuous cardiac output determinations 
relates to the inaccuracies of pulse oximetry and 
venous oximetry, as discussed earlier in this paper. 
Either Vo 2 or \fc02 can be used to calculate Fick 
cardiac output, and the use of Vcoz for continuous 
cardiac output measurements has been described. 191 
Relatively good correlations between continuous 
Fick cardiac output determinations and intermittent 
thermodilution cardiac output determinations have 
been reported (r > 0.8). I86191 

Gastric Tonometry 

Gastric tonometry has been suggested as a mini- 
mally invasive method to assess tissue oxygenation 
in critically ill patients. 192 19S Perfusion of the gut is 
affected early in the course of systemic hypoxia. 
Theoretically, tissue hypoxia should result in a de- 
crease in gastric intramucosal pH (pH,). Normal 
pH, is 7.38 ± 0.03. I94 

The gastric tonometer consists of a nasogastric 
tube with a distal C0 2 -permeable balloon (Fig. 
24) iw-198 The ba iioon is filled with 2.5-3 mL of 
physiologic saline solution. An equilibration time 
of 1-2 hours is allowed for C0 2 in the gastric lu- 
men to equilibrate with the saline inside the bal- 
loon. After discarding 1-1.5 mL of aspirate from 
the gastric tube (dead volume), the remaining 1-1.5 
mL is analyzed for Pco: using a blood gas analyz- 
er. A simultaneous arterial blood sample is an- 
alyzed to determine HCO, . Gastric pH, is then cal- 
culated from the Henderson-Hasselbach equation, 
using the Pcc>2 of the saline from the gastric bal- 
loon and the HC0 3 of arterial blood. 

Fig. 24. Schematic illustration of gastric tonometer in the 

The use of gastric tonometry to evaluate tissue 
oxygenation is relatively new. Although much more 
work is needed to evaluate its clinical usefulness, 192 
a recent study of 22 critically ill patients reported 
that pH, < 7.32 had a sensitivity of 89% and a spec- 
ificity of 77% as an early predictor of mortality. 197 

Monitoring in Perspective 

How much monitoring is needed? This is an im- 
portant question for both clinicians and admin- 
istrators. Clinicians often want to monitor every- 
thing possible, with a 'more-is-better' attitude. On 
the other hand, administrators become justifiably 
concerned with the costs associated with mon- 
itoring. It is our clinical impression that monitoring 
can be excessive at times, and may distract from, 
rather than complement, patient care. Critical care 
monitoring has become widespread, without ap- 
propriately designed studies to evaluate its impact 
on outcome. In many cases, monitoring has be- 
come the standard of care. As stated by Keats," 5 
"Unfortunately, making a monitor part of the stan- 
dard of care guarantees that the experiment nec- 
essary to document its benefits will never be carried 
out because we have ruled out a control group." 

With some monitors, clinicians do not thorough- 
ly understand the physiologic significance and clin- 
ical limitations of the numbers displayed by the 
monitor. There is a tendency for many clinicians to 
take the displayed results at face value, with little 
understanding of how the data were collected, how 
the data were processed inside the "black box," and 
what restrictions limit the usefulness of the results. 

The presence of many monitors at the bedside 
can be very distracting to clinicians. Many mon- 
itoring systems tend to beep, buzz, and blink con- 
stantly. Sometimes we seem to spend more time 
caring for monitors than caring for patients. Un- 
fortunately, the monitor is often silenced or ig- 
nored, at which point it becomes virtually worth- 
less. If the incidence of false-positive alarms is 5% 
in a given monitor, then the probability of at least 1 
monitor's giving a false-positive alarm at any time 
will be 64% if 20 monitors are used. 1 This becomes 
confusing for clinicians, who are faced with the de- 
cision to either ignore the alarm or to treat the pa- 
tient. Bentt et al 199 found that a pulse oximeter 
alarm was present up to 47% of the time (28 min/h) 




in a 10-bed surgical ICU, and that many of these 
were false alarms and required no intervention. 
During anesthesia monitoring, Kestin et al 200 found 
that 75% of all alarms that sounded were spurious 
and that only 3% indicated risk to the patient. Bow- 
ton et al 72 found that use of pulse oximetry outside 
the ICU had little effect on physician-directed res- 
piratory care. 

The financial impact of monitoring is difficult to 
assess. For noninvasive monitors such as pulse oxi- 
meters, no prospective randomized controlled trials 
have evaluated cost-effectiveness. Using a his- 
torical control and a relatively short evaluation pe- 
riod (2 months), Kellerman et al™ reported a re- 
duction in the number of ABGs ordered in an 
Emergency Department following the introduction 
of pulse oximetry. However, others have shown an 
increased number of ABGs when pulse oximetry 
was used. 28 For mixed venous oximetry, Orlando 
reported a savings of $75/patient. 2ul Pearson et 
al, 202 however, found that monitoring mixed ve- 
nous saturation added significantly to the cost in- 
curred with a routine pulmonary artery catheter. 
Jastremski et al 203 also found that mixed venous ox- 
imetry may not be cost-effective, particularly in a 
fixed-payment reimbursement system. 

Monitoring of oxygenation is often useful in the 
care of critically ill patients. However, monitoring 
should not be done just because it is technically 
feasible. Technical capability must be balanced 
against clinical usefulness and cost-effectiveness. 
The decision to monitor, like any other clinical de- 
cision, should be based on therapeutic objectives. 

When technology is master, 
We shall reach disaster faster. 

Hein & Grooks 2 " 4 

1. Hess D. Noninvasive monitoring in respiratory care — 
present, past, and future: an overview. Respir Care 

2. Hess D. Noninvasive respiratory monitoring during ven- 
tilatory support. Crit Care Nurs Clin North Am 1991:3: 

3. Jubran A, Tobin MJ. Noninvasive oxygen monitoring. 
In: Cheng EY, Lund N, eds. Tissue oxygenation. Probl 
Crit Care 1992:6:394-407. 

4. Ortiz CR. Lund N. Invasive oxygen monitoring. In: 
Cheng EY. Lund N. eds. Tissue oxygenation. Probl Crit 
Care 1992;6:408-422. 

5. Hoyt JW. Controversies in oxygen monitoring. In: 
Cheng EY, Lund N, eds. Tissue Oxygenation. Probl 
Crit Care 1992;6:443-450. 

6. Clark LC Jr. Clark EW. A personalized history of the 
Clark oxygen electrode. Int Anesthesiol Clin 1987:25 
(3): 1-29. 

7. AARC Clinical practice guideline: sampling for arterial 
blood gas analysis. Respir Care 1992:37:913-917. 

8. Raffin TA. Indications for arterial blood gas analysis. 
Ann Intern Med 1986;105:390-398. 

9. Moran RF. Assessment of quality control of blood gas/ 
pH analyzer performance. Respir Care 1981:26:538- 

10. Elser RC. Quality control of blood gas analysis: a re- 
view. Respir Care 1986;31:807-816. 

11. Mahoney JJ, Harvey JA. Wong RJ, Van Kessel AL. 
Changes in oxygen measurements when whole blood is 
stored in iced plastic or glass syringes. Clin Chem 1991 ; 

12. National Committee for Clinical Laboratory Standards 
(NCCLS) Document C27-T. Blood gas preanalytic con- 
siderations: specimen collection, calibration, and con- 
trols: tentative guideline. Villanova PA: NCCLS. 1989. 

13. National Committee for Clinical Laboratory Standards 
(NCCLS) Document Hll-A. Percutaneous collection of 
arterial blood for laboratory analysis: approved stan- 
dard. Villanova PA: NCCLS, 1985. 

14. National Committee for Clinical Laboratory Standards 
(NCCLS) Document C21-T. Performance character- 
istics for devices measuring Po: and Pco: in blood sam- 
ples: tentative standard. Villanova PA: NCCLS. 1989. 

15. National Committee for Clinical Laboratory Standards 
(NCCLS) Document M29-T. Protection of laboratory 
workers from infectious disease transmitted by blood, 
body fluids, and tissue: tentative standard. Villanova 

16. National Committee for Clinical Laboratory Standards 
(NCCLS) Document H24-P. Additives to blood collec- 
tion devices: heparin. Villanova PA: NCCLS. 1985. 

17. Hansen JE, Simmons DH. A systematic error in the de- 
termination of blood Pco;- Am Rev Respir Dis 1977: 

18. Fleisher M, Gladstone M. Crystal D, Schwartz MK. 
Two whole-blood multianalyte analyzers evaluated. 
Clin Chem 1989;35:1532-1535. 

19. Bryant MT. Gases stored in disposable syringes: a study 
of changes in their concentrations. Anaesthesia 1977:32: 

20. Evers W, Racz GB, Levy AA. A comparative study of 
plastic (polypropylene) and glass syringes in blood-gas 
analysis. Anesth Analg 1972:51:92-97. 

21. Bageant RA. Variations in arterial blood gas measure- 
ments due to sampling techniques. Respir Care 1975; 

22. Biswas CK, Ramos JM. Agroyannis B. Kerr DN. Blood 
gas analysis: effect of air bubbles in syringe and delay 
in estimation. Br Med J Clin Res Ed 1982:284:923-927. 




23. Scheinhorn DJ, Angelillo VA. Heparin sodium and ar- 
terial blood gas analysis (letter). Chest 1978;73:244-245. 

24. Goodwin NM, Schreiber MT. Effects of anticoagulants 
on acid-base and blood gas estimations. Crit Care Med 

25. Madiedo G. Sciacca R, Hause L. Air bubbles and tem- 
perature effect on blood gas analysis. J Clin Pathol 

26. Mueller RG, Lang GE. Beam JM. Bubbles in samples 
for blood gas determinations: a potential source of error. 
Am J Clin Pathol 1976:65:242-249. 

27. Hansen JE. Participant responses to blood gas pro- 
ficiency testing reports. Chest 1992:101: 1240-1244. 

28. Muakkassa FF, Rutledge R, Fakhry SM, Meyer AA, 
Sheldon GF. ABGs and arterial lines: the relationship to 
unnecessarily drawn arterial blood gas samples. J Trau- 
ma 1990:30:1087-1095. 

29. Smoller BR. Kruskall MS. Phlebotomy for diagnostic 
laboratory tests in adults: pattern of use and effect on 
transfusion requirements. N Engl J Med 1986:314:1233- 

30. Beasley KE. Darin JM, Durbin CG Jr. The effect of res- 
piratory care department management of a blood gas an- 
alyzer on the appropriateness of arterial blood gas utili- 
zation. RespirCare 1992;37:343-347. 

31. Hess D, Agarwal NN. Variability of blood gases, pulse 
oximeter saturation, and end-tidal carbon dioxide pres- 
sure in stable, mechanically ventilated trauma patients. J 
ClinMonit 1992:8:111-115. 

32. Thorson SH, Marini JJ, Pierson DJ, Hudson LD. Vari- 
ability of arterial blood gas values in stable patients in 
thelCU. Chest 1983;84:14-18. 

33. National Committee for Clinical Laboratory Standards 
(NCCLS) Document C25-P. Oxygen content, hemo- 
globin oxygen, "saturation." and related quantities in 
blood: terminology, measurement, and reporting. Villa- 
nova PA: NCCLS. 

34. Hess D, Elser RC, Agarwal NN. The effects on the pul- 
monary shunt value of using measured versus calculated 
hemoglobin oxygen saturation and of correcting for the 
presence of carboxyhemoglobin and methemoglobin. 
RespirCare 1984;29:1001-1005. 

35. Hess D, Maxwell C, Shefet D. Determination of intra- 
pulmonary shunt: comparison of an estimated shunt 
equation and a modified equation with the classic equa- 
tion. RespirCare 1987;32:268-273. 

36. Cohn JD, Engler PE. Shunt effect of carboxyhemo- 
globin. Crit Care Med 1979;7:54-58. 

37. Cane RD. Shapiro BA, Harrison RA, Steiner MC, Kav- 
anaugh J. Minimizing errors in intrapulmonary shunt 
calculations. Crit Care Med 1980;8:294-297. 

38. Zijlstra WG, Buursma A. Meeuwsen-van der Roest WP. 
Absorption spectra of human fetal and adult oxyhemo- 
globin, deoxyhemoglobin, carboxyhemoglobin, and met- 
hemoglobin. Clin Chem 1 99 1 ;37: 1 633- 1 638. 

39. Mahoney JJ. Wong RJ. Vreman HJ, Stevenson DK. Fe- 
tal hemoglobin of transfused neonates and spectro 

photometric measurements of oxyhemoglobin and car- 
boxyhemoglobin. J Clin Monit 1991,7: 154-160. 

40. Cornelissen PJ. van Woensel CL, van Oel WC, de Jong 
PA. Correction factors for hemoglobin derivatives in fe- 
tal blood, as measured with the IL-282 CO-oximeter 
(letter). Clin Chem 1983:29:1555-1556. 

4 1 . Sehgal LR, Sehgal HL, Rosen AL, Gould SA. Moss GS. 
Effect of intralipid on measurements of total hemo- 
globin and oxyhemoglobin in whole blood. Crit Care 
Med 1984;12:907-909. 

42. Chernow B. The bedside laboratory: a critical step for- 
ward in ICU care. Chest 1990:97(5, Suppl):183S-184S. 

43. Misiano DR. Meyerhoff ME, Collison ME. Current and 
future directions in the technology relating to bedside 
testing of critically ill patients. Chest 1990:97(5. Suppl): 

44. Zaloga GP. Evaluation of bedside testing options for the 
critical care unit. Chest 1990:97(5, Suppl):185S-190S. 

45. Belsey R, Baer D, Sewell D. Laboratory test analysis 
near the patient: opportunities for improved clinical di- 
agnosis and management. JAMA 1986;255:775-786. 

46. Belsey R, Morrison JI, Whitlow KJ, Baer DM, Nelson 
S, Hardwick DF. Managing bedside glucose testing in 
the hospital. JAMA 1987;258:1634-1638. 

47. Bashein G. Greydanus WK, Kenny MA. Evaluation of a 
blood gas and chemistry monitor for use during surgery. 
Anesthesiology 1989:70:123-127. 

48. Belsey R, Goitein RK, Baer DM. Evaluation of a labor- 
atory system intended for use in physicians' offices: I. 
reliability of results produced by trained laboratory tech- 
nologists. JAMA 1987;258:353-356. 

49. Belsey R, Vandenbark M, Goitein RK, Baer DM. Eval- 
uation of a laboratory system intended for use in phy- 
sicians' offices: II. reliability of results produced by 
health care workers without formal or professional la- 
boratory training. JAMA 1987:258:357-361. 

50. Shapiro BA. In-vivo monitoring of arterial blood gases 
and pH. RespirCare 1992:37:165-169. 

51. Peterson JI. Vurek GG. Fiber-optic sensors for bio- 
medical applications. Science 1984;224:123-127. 

52. Miller WW, Gehrich JL, Hansmann DR, Yafuso M. 
Continuous in vivo monitoring of blood gases. Lab Med 

53. Barker SJ. Tremper KK. Intra-arterial oxygen tension 
monitoring. Int Anesthesiol Clin 1987:25(3): 199-208. 

54. Bratanow N, Polk K. Bland R, Kram HB. Lee TS. Shoe- 
maker WC. Continuous polarographic monitoring of in- 
tra-arterial oxygen in the perioperative period. Crit Care 
Med 1985:13:859-860. 

55. Arai T, Hatano Y, Komatsu K, Takada T, Miyake C, 
Harioka T, et al. Real-time analysis of the change in ar- 
terial oxygen tension during endotracheal suction with a 
fiberoptic bronchoscope. Crit Care Med 1985:13:855- 

56. Green GE, Hassell KT. Mahutte CK. Comparison of ar- 
terial blood gas with continuous intra-arterial and trans- 
cutaneous Po: sensors in adult critically ill patients. Crit 
Care Med 1987:15:491-494. 




57. Malalis L, Bhat R. Vidyasagar D. Comparison of intra- 
vascular Po; with transcutaneous and PjO: values. Crit 
Care Med 1983:11:110-113. 

58. Conway M, Durbin GM, Ingram D. Mcintosh N, Parker 
D, Reynolds OR, et al. Continuous monitoring of arteri- 
al oxygen tension using a catheter-tip polarographic 
electrode in infants. Pediatrics 1976:57:244-250. 

59. Harris TR. Nugent M. Continuous arterial oxygen ten- 
sion monitoring in the newborn infant. J Pediatr 1973; 

60. Gehrich JL, Lubbers DW. Opitz N, Hansmann DR. 
Miller WW, Tusa JK, et al. Optical fluorescence and its 
application to an intravascular blood gas monitoring 
system. IEEE Trans Biomed Eng 1986:33:1 17-132. 

61. Angel SM. Optrodes: chemically selective fiber-optic 
sensors. Spectroscopy 1987:2:38-46. 

62. Mook GA, Osypka P. Sturm RE, Wood EH. Fibre optic 
reflection photometry on blood. Cardiovasc Res 1968; 

63. Opitz N, Lubbers DW. Theory and development of fluo- 
rescence-based optochemical oxygen sensors: oxygen 
optodes. Int Anesthesiol Clin 1 987;25( 3 ): 1 77- 1 97. 

64. Shapiro BA. Cane RD. Chomka CM, Bandala LE, Pe- 
ruzzi WT. Preliminary evaluation of an intra-arterial 
blood gas system in dogs and humans. Crit Care Med 

65. Mahutte CK, Sassoon CS. Muro JR. Hansmann DR. 
Maxwell TP, Miller WW, et al. Progress in the develop- 
ment of a fluorescent intravascular blood gas system in 
man. J Clin Monit 1990:6:147-157. 

66. Miller WW. Yafuso M. Yan CF, Hui HK. Arick S. Per- 
formance of an in-vivo, continuous blood-gas monitor 
with disposable probe. Clin Chem 1987;33:1538-1542. 

67. Barker SJ, Tremper KK, Hyatt J, Zaccari J. Heitzmann 
HA. Holman BM. et al. Continuous fiberoptic arterial 
oxygen tension measurements in dogs. J Clin Monit 

68. Severinghaus JW. Kelleher JF. Recent developments in 
pulse oximetry. Anesthesiology 1992:76:1018-1038. 

69. Aughey K. Hess D. Eitel D. Bleecher K. Cooley M. Og- 
den C, et al. An evaluation of pulse oximetry in pre- 
hospital care. Ann Emerg Med 1991;20:887-891. 

70. Kellerman AL, Cofer CA, Joseph S, Hackman BB. Im- 
pact of portable pulse oximetry on arterial blood gas test 
ordering in an urban emergency department. Ann Emerg 
Med 1991:20:130-134. 

71. Hay WW Jr. Thilo E. Curlander JB. Pulse oximetry in 
neonatal medicine. Clin Perinatol 1991:18:441-472. 

72. Bowton DL, Scuderi PE, Harris L, Haponik EF. Pulse 
oximetry monitoring outside the intensive care unit: 
progress or problem? Ann Intern Med 1991:115:450- 

73. Schnapp LM. Cohen NH. Pulse oximetry: uses and 
abuses. Chest 1990:98: 1244-1250. 

74. Tremper KK. Barker SJ. Pulse oximetry. Anesthes- 
iology 1989:70:98-108. 

75. Kelleher JF. Pulse oximetry. J Clin Monit 1989:5:37-62. 

76. Alexander CM. Teller LE, Gross JB. Principles of pulse 
oximetry: theoretical and practical considerations. 
AnesthAnalg 1989:68:368-376. 

77. Welch JP. DeCesare R, Hess D. Pulse oximetry: instru- 
mentation and clinical applications. Respir Care 1990; 

78. Palve H. Reflection and transmission pulse oximetry 
during compromised peripheral perfusion. J Clin Monit 

79. Mendelson Y. McGinn MJ. Skin reflectance pulse ox- 
imetry: in vivo measurements from the forearm and 
calf. J Clin Monit 1991;7:7-12. 

80. Palve H. Comparison of reflection and transmission 
pulse oximetry after open-heart surgery. Crit Care Med 

81. Ralston AC. Webb RK, Runciman WB. Potential errors 
in pulse oximetry. I: pulse oximeter evaluation. Anae- 
thesia 1991:46:202-206. 

82. Webb RK. Ralston AC, Runciman WB. Potential errors 
in pulse oximetry. II: effects of changes in saturation 
and signal quality. Anaesthesia 1 99 1 ;46:207-2 1 2. 

83. Ralston AC. Webb RK. Runciman WB. Potential errors 
in pulse oximetry. Ill: effects of interference, dyes, dys- 
haemoglobins and other pigments. Anaesthesia 1991:46: 

84. Nickerson BG. Sarkisian C. Tremper K. Bias and pre- 
cision of pulse oximeters and arterial oximeters. Chest 

85. Choe H, Tashiro C. Fukumitsu K. Yagi M, Yoshiya 1. 
Comparison of recorded values from six pulse oxi- 
meters. Crit Care Med 1 989; 1 7:678-68 1 . 

86. Taylor MB. Whitwam JG. The accuracy of pulse ox- 
imeters: a comparative clinical evaluation of five pulse 
oximeters. Anaesthesia 1988:43:229-232. 

87. Hannhart B. Michalski H. Delorme N. Chapparo G, 
Pulu JM. Reliability of six pulse oximeters in chronic 
obstructive pulmonary disease. Chest 1991:99:842-846. 

88. Severinghaus JW. Naifeh KH. Koh SO. Errors in 14 
pulse oximeters during profound hypoxia. J Clin Monit 

89. Severinghaus JW, Naifeh KH. Accuracy of response of 
six pulse oximeters to profound hypoxia. Anesthesi- 
ology 1987:67:551-558. 

90. Hannhart B. Haberer JP, Saunier C, Laxenaire MC. Ac- 
curacy and precision of fourteen pulse oximeters. Eur 
Respir J 1991;4:115-119. 

91. Kelleher JF. Ruff RH. The penumbra effect: vasomo- 
tion-dependent pulse oximeter artifact due to probe mal- 
position. Anesthesiology 1989:71:787-791. 

92. Serpell MG. Children's fingers and spurious pulse ox- 
imetry (letter). Anaesthesia 1991:46:702-703. 

93. Vegfors M. Lennmarken C. Carboxyhaemoglobinaemia 
and pulse oximetry. Br J Anaesth 1991:66:625-626. 

94. Barker SJ. Tremper KK, Hyatt J. Effects of methe- 
moglobinemia on pulse oximetry and mixed venous ox- 
imetry. Anesthesiology 1989:70:1 12-117. 








95. Barker SJ. Tremper KK. The effect of carbon monoxide 
inhalation on pulse oximetry and transcutaneous Poj. 
Anesthesiology 1987;66:677-679. 
Raemer DB, Elliott WR, Topulos GP, Philip JH. The 
theoretical effect of carboxyhemoglobin on the pulse 
oximeter. J Clin Monit 1989;5:246-249. 
Harris AP, Sendak JM. Donham RT, Thomas M. Dun- 
can D. Absorption characteristics of human fetal hemo- 
globin at wavelengths used in pulse oximetry. J Clin 
Monit 1988:4:175-177. 

Scott DM, Cooper MG. Spurious pulse oximetry with 
intrauterine methylene blue injection. Anaesth Intensive 
Care 1991:19:267-268. 

Scheller MS. Unger RJ, Reiner MJ. Effects of intra- 
venously administered dyes on pulse oximeter readings. 
Anesthesiology 1986;65:550-552. 
Cote CJ. Goldstein EA. Fuchsman WH, Hoaglin DC. 
The effect of nail polish on pulse oximetry. Anesth 
Analg 1989;67:683-686. 

101. Veyckemans F. Baele P, Guillaume JE. Willems E, 
Robert A. Clerbaux T. Hyperbilirubinemia does not in- 
terfere with hemoglobin saturation measured by pulse 
oximetry. Anesthesiology 1989;70:118-122. 

102. Chelluri L, Snyder JV. Bird JR. Accuracy of pulse ox- 
imetry in patients with hyperbilirubinemia. Respir Care 

103. Zeballos RJ, Weisman IM. Reliability of noninvasive 
oximetry in black subjects during exercise and hypoxia. 
Am Rev Respir Dis 1991;144:1240-1244. 

104. Jubran A, Tobin MJ. Reliability of pulse oximetry in ti- 
trating supplemental oxygen therapy in ventilator- 
dependent patients. Chest 1990;97:1420-1425. 
Cahan C. Decker MJ, Hoekje PL, Strohl KP. Agreement 
between noninvasive oximetric values for oxygen sat- 
uration. Chest 1990;97:814-819. 

Severinghaus JW. Koh SO. Effect of anemia on pulse 
oximeter accuracy at low saturation. J Clin Monit 1990; 

107. Lee S, Tremper KK. Barker SJ. Effects of anemia on 
pulse oximetry and continuous mixed venous hemo- 
globin saturation monitoring in dogs. Anesthesiology 

108. Sami HM, Kleinman BS, Lonchyna VA. Central venous 
pulsations associated with a falsely low oxygen satura- 
tion measured by pulse oximetry. J Clin Monit 1991:7: 

109. Wilkins CJ, Moores M, Hanning CD. Comparison of 
pulse oximeters: effects of vasoconstriction and venous 
engorgement. Br J Anaesth 1989;62:439-444. 

110. Stewart KG. Rowbottom SJ. Inaccuracy of pulse oxim- 
etry in patients with severe tricuspid regurgitation. An- 
aesthesia 1991;46:668-670. 

111. Chemello PD. Nelson SR, Wolford LM. Finger injury 
resulting from pulse oximeter probe during orthognathic 
surgery. Oral Surg Oral Med Oral Pathol 1990:69:161- 

112. Murphy KG, Secunda JA, Rockoff MA. Severe burns 
from a pulse oximeter. Anesthesiology 1990;73:350-352. 










Stogner SW, Owens MW. Baethge BA. Cutaneous ne- 
crosis and pulse oximetry. Cutis 1991:48:235-237. 
Bashein G, Syrory G. Burns associated with pulse oxim- 
etry during magnetic resonance imaging. Anesthesi- 
ology 1991:75:382-383. 

Keats AS. Anesthesia mortality in perspective. Anesth 
Analg 1990:71:113-119. 

Freund PR. Overand PT. Cooper J, Jacobson L. Bosse 
S, Walker B. et al. A prospective study of intraoperative 
pulse oximetry failure. J Clin Monit 1991:7:253-258. 
Cheng EY, Lauer KK, Stommel KA. Guenther NR. 
Evaluation of the palmar circulation by pulse oximetry. 
J Clin Monit 1989:5:1-3. 

Pillow K. Herrick IA. Pulse oximetry compared with 
Doppler ultrasound for assessment of collateral blood 
flow to the hand. Anaesthesia 1991:46:388-390. 
Talke PO. Measurement of systolic blood pressure using 
pulse oximetry during helicopter flight. Crit Care Med 

Chawla R, Kumarvel V. Girdhar KK, Sethi AK. Indra- 
yan A, Bhattacharya A. Can pulse oximetry be used to 
measure systolic blood pressure? Anesth Analg 1992: 

Talke P, Nichols RJ Jr. Traber DL. Does measurement 
of systolic blood pressure with a pulse oximeter cor- 
relate with conventional methods'? J Clin Monit 1990; 

Martin RJ. Transcutaneous monitoring: instrumentation 
and clinical applications. Respir Care 1990;35:577-583. 
Lubbers DW. Theory and development of transcutane- 
ous oxygen pressure measurement. Int Anesthesiol Clin 

Tremper KK, Barker SJ. Transcutaneous oxygen meas- 
urement: experimental studies and adult applications. 
Int Anesthesiol Clin 1987;25(3):67-96. 
Kesten S, Chapman KR. Rebuck AS. Response char- 
acteristics of a dual transcutaneous oxygen/carbon diox- 
ide monitoring system. Chest 1 99 1 ;99: 1 2 1 1 - 1 2 1 5 . 
Hasibeder W, Haisjackl M, Sparr H. Klaunzer S. Hor- 
man C. Salak N, et al. Factors influencing trans- 
cutaneous oxygen and carbon dioxide measurements in 
adult intensive care patients. Intensive Care Med 1991: 

Palmisano BW. Severinghaus JW. Transcutaneous 
PCO2: a multicenter study of accuracy. J Clin Monit 

Rooke TW. Osmundson PJ. Variability and re- 
producibility of transcutaneous oxygen tension measure- 
ments in the assessment of peripheral vascular disease. 
Angiology 1989;40:695-700. 

Tremper KK. Waxman K. Shoemaker WC. Effects of 
hypoxia and shock on transcutaneous Po: values in 
dogs. Crit Care Med 1979;7:526-531. 
Dronen SC. Maningas PA, Foutch R. Transcutaneous 
oxygen tension measurements during graded hemor- 
rhage and reinfusion. Ann Emerg Med 1985:14:534- 




131. Gottrup F, Gellett S, Kirkegaard L, Hansen ES, Johan- 
sen G. Effect of hemorrhage and resuscitation on sub- 
cutaneous, conjunctival, and transcutaneous oxygen ten- 
sion in relation to hemodynamic variables. Crit Care 
Med 1989;17:904-907. 

132. Tremper KK, Shoemaker WC. Transcutaneous oxygen 
monitoring of critically ill adults, with and without low 
flow shock. Crit Care Med 1981:9:706-709. 

133. Tremper KK. Waxman K. Bowman R. Shoemaker WC. 
Continuous transcutaneous oxygen monitoring during 
respiratory failure, cardiac decompensation, cardiac ar- 
rest, and CPR: transcutaneous oxygen monitoring dur- 
ing arrest and CPR. Crit Care Med 1980:8:377-381. 

134. Abraham E, Smith M, Silver L. Continuous monitoring 
of critically ill patients with transcutaneous oxygen and 
carbon dioxide and conjunctival oxygen sensors. Ann 
EmergMed 1984; 13( 1 1 ): 102 1-1026. 

135. Waxman K, Sadler R. Eisner ME, Applebaum R, Trem- 
per KK. Mason GR. Transcutaneous oxygen monitoring 
of emergency department patients. Am J Surg 1983:146: 

136. Abraham E, Smith M, Silver L. Conjunctival and trans- 
cutaneous oxygen monitoring during cardiac arrest and 
cardiopulmonary resuscitation. Crit Care Med 1984:12: 

137. Abraham E, Ehrlich H. Conjunctival and transcutaneous 
oxygen monitoring during resuscitation. Ann Emerg 
Med 1984;13:287-289. 

138. Dantzker DR. Oxygen transport and utilization. Respir 
Care 1988;33:874-880. 

139. Nelson LD. Rutherford EJ. Monitoring mixed venous 
oxygen. RespirCare 1992;37:154-164. 

140. Kasnitz P. Druger GL. Yorra F. Simmons DH. Mixed 
venous oxygen tension and hyperlacticemia: survival in 
severe cardiopulmonary disease. JAMA 1976:236:570- 

141. Suter PM, Lindauer JM. Fairley HB, Schlobohm RM. 
Errors in data derived from pulmonary artery blood gas 
values. Crit Care Med 1975:3:175-181. 

142. Mihm F. Feeley TW. Rosenthal M, Raffin TA. The lack 
of effect of variable blood withdrawal rates on the meas- 
urement of mixed venous oxygen saturation. Chest 

143. Nelson RD. Wilkins RL, Jacobsen WK, Sheldon RL. 
Supranormal PvO? in the presence of tissue hypoxia: a 
case report. RespirCare 1983:28:191-194. 

144. Nelson LD. Continuous venous oximetry in surgical pa- 
tients. Ann Surg 1986;203:329-333. 

145. Tahvanainen J. Meretoja O. Nikki P. Can central venous 
blood replace mixed venous blood samples? Crit Care 
Med 1982;10:758-761. 

146. Schweiss JF. Mixed venous hemoglobin saturation: the- 
ory and application. Int Anesthesiol Clin 1987:25(3): 

147. Paulus DA. Invasive monitoring of respiratory gas ex- 
change: continuous measurement of mixed venous oxy- 
gen saturation. RespirCare 1987;32:535-543. 








Fahey PJ, Harris K. Vanderwarf C. Clinical experience 
with continuous monitoring of mixed venous oxygen 
saturation in respiratory failure. Chest 1984;86:748-752. 
Baraka A, Baroody M, Haroun S. Nawfal M. Dabbous 
A. Sibai A, et al. Continuous venous oximetry during 
cardiopulmonary bypass: influence of temperature 
changes, perfusion flow, and hematocrit levels. J Car- 
diothorac Anesth 1990;4:35-38. 

Howdieshell TR. Sussman A, Dipiro J, McCarten M, 
Mansberger A Jr. Reliability of in vivo mixed venous 
oximetry during experimental hypertriglyceridemia. Crit 
Care Med 1992;20:999-1004. 

Rouby JJ, Poete P, Bodin L, Bourgeois JL, Arthaud M. 
Viars P. Three mixed venous saturation catheters in pa- 
tients with circulatory shock and respiratory failure. 
Chest 1990:98:954-958. 

Baele PL, McMichan JC, Marsh HM, Sill JC, Southorn 
PA. Continuous monitoring of mixed venous oxygen 
saturation in critically ill patients. Anesth Analg 1982; 

Waller JL, Kaplan JA, Bauman DI. Craver JM. Clinical 
evaluation of a new fiberoptic catheter oximeter during 
cardiac surgery. Anesth Analg 1982;61:676-679. 
Gettinger A. DeTraglia MC. Glass DD. In vivo com- 
parison of two mixed venous saturation catheters. Anes- 
thesiology 1987;66:373-375. 

Karis JH, Lumb PD. Clinical evaluation of the Edwards 
Laboratories and Oximetrix mixed venous oxygen sat- 
uration catheters. J Cardiothorac Anesth 1988:2:440-444. 
Divertie MB. McMichan JC. Continuous monitoring of 
mixed venous oxygen saturation. Chest 1984:85:423-428. 
Reinhart K, Moser N, Rudolph T, Bredle D, Specht M, 
Gramm HJ, et al. Accuracy of two mixed venous satura- 
tion catheters during long-term use in critically ill pa- 
tients. Anesthesiology 1988:69:769-773. 
Hecker BR. Brown DL, Wilson D. A comparison of two 
pulmonary artery mixed venous oxygen saturation cath- 
eters during the changing conditions of cardiac surgery. 
J Cardiothorac Anesth 1989;3:269-275. 
Schranz D. Schmitt S. Oelert H, Schmid F. Huth R. 
Zimmer B, et al. Continuous monitoring of mixed ve- 
nous oxygen saturation in infants after cardiac surgery. 
Intensive Care Med 1989.15:228-232. 
Van Woerkens EC. Trouwborst A. Tenbrinck R. Accu- 
racy of a mixed venous saturation catheter during acute- 
ly induced changes in hematocrit in humans. Crit Care 
Med 1991;19:1025-1029. 

Vaughn S. Puri VK. Cardiac output changes and con- 
tinuous mixed venous oxygen saturation measurement 
in the critically ill. Crit Care Med 1988;16:495-498. 
Bland JM. Altman DG. Statistical methods for assessing 
agreement between two methods of clinical measure- 
ment. Lancet 1986:1:307-310. 

Scuderi PE. Bowton DL. Eredith JW, Harris LC. Evans 
JB, Anderson RL. A comparison of three pulmonary ar- 
tery oximetry catheters in intensive care patients. Chest 




164. Rasanen J, Downs JB, Hodges MR. Continuous mon- 
itoring of gas exchange and oxygen use with dual oxim- 
etry. J Clin Anesth 1988;1:3-8. 

165. Zaune U, Knarr C, Kruselmann M, Pauli MHF. Boeden 
G, Martin E. Value and accuracy of dual oximetry dur- 
ing pulmonary resections. J Cardiothorac Anesth 1990; 

166. Sumimoto T, Takayama Y, Iwasaka T, Sugiura T, Ta- 
keuchi M, Hasegawa T, et al. Mixed venous oxygen sat- 
uration as a guide to tissue oxygenation and prognosis 
in patients with acute myocardial infarction. Am Heart J 

167. Magilligan DJ Jr. Teasdall R, Eisinminger R. Peterson 
E. Mixed venous oxygen saturation as a predictor of car- 
diac output in the postoperative cardiac surgical patient. 
Ann Thorac Surg 1987;44:260-262. 

168. Norfleet EA. Watson CB. Continuous mixed venous 
oxygen saturation measurement: a significant advance 
in hemodynamic monitoring? J Clin Monit 1 985; 1:245- 

169. Rasanen J, Downs JB, Malec DJ, DeHaven B, Seidman 
P. Estimation of oxygen utilization by dual oximetry. 
Ann Surg 1987;206:621-623. 

170. Rasanen J, Downs JB. DeHaven B. Titration of continu- 
ous positive airway pressure by real-time dual oximetry. 
Chest 1987;92:853-856. 

171. Rasanen J, Downs JB, Malec DJ, DeHaven B, Garner 
WL. Real-time continuous estimation of gas exchange 
by dual oximetry. Intensive Care Med 1988:14:1 18-122. 

172. Rasanen J, Downs JB, Malec DJ, Oates K. Oxygen ten- 
sions and oxyhemoglobin saturations in the assessment 
of pulmonary gas exchange. Crit Care Med 1987:15: 

173. Liggett SB, St John RE, Lefrak SS. Determination of 
resting energy expenditure utilizing the thermodilution 
pulmonary artery catheter. Chest 1987;91:562-566. 

174. Liggett SB, Renfro AD. Energy expenditures of me- 
chanically ventilated nonsurgical patients. Chest 1990: 

175. Brandi LS, Grana M, Mazzanti T, Giunta F, Natali A, 
Ferrannini E. Energy expenditure and gas exchange 
measurements in postoperative patients: thermodilution 
versus indirect calorimetry. Crit Care Med 1992:20: 

176. Bartlett RH, Deckert RE. Oxygen kinetics: pitfalls in 
clinical research. J Crit Care 1990;5:77-80. 

177. Myburgh JA, Webb RK, Worthley LI. Ventilation/ 
perfusion indices do not correlate with the difference be- 
tween oxygen consumption measured by the Fick prin- 
ciple and metabolic monitoring systems in critically ill 
patients. Crit Care Med 1992;20:479-482. 

178. Smithies MN, Royston B, Makita K, Konieczko K, 
Nunn JF. Comparison of oxygen consumption measure- 
ments: indirect calorimetry versus the reversed Fick 
method. Crit Care Med 1991;19:1401-1406. 

179. Light RB. Intrapulmonary oxygen consumption in ex- 
perimental pneumococcal pneumonia. J Appl Physiol 

180. Archie JP Jr. Mathematic coupling of data: a common 
source of error. Ann Surg 1981:193:296-303. 

181. Moreno LF, Stratton HH. Newell JC. Fenstel PJ. Math- 
ematical coupling of data: correction of a common error 
for linear calculations. J Appl Physiol 1986:60:335-343. 

182. Stratton HH. Fenstel PJ. Newell JC. Regression of cal- 
culated variables in the presence of shared measurement 
error. J Appl Physiol 1987:62:2083-2093. 

183. Ronco JJ, Phang PT, Walley KR, Wiggs B. Fenwick JC. 
Russell JA. Oxygen consumption is independent of 
changes in oxygen delivery in severe adult respiratory 
distress syndrome. Am Rev Respir Dis 1991:143:1267- 

184. Branson RD. The measurement of energy expenditure: 
instrumentation, practical considerations, and clinical 
application. Respir Care 1990;35:640-659. 

1 85. Hess D, Mundorff J. Assessment of metabolic and nutri- 
tional status. In: Pierson DJ, Kacmarek RM. eds. Foun- 
dations of respiratory care. New York: Churchill Living- 
stone, 1992:541-554. 

186. Carpenter JP, Nair S. Staw I. Cardiac output determina- 
tion: thermodilution versus a new computerized Fick 
method. Crit Care Med 1985;13:576-579. 

187. Quinn TJ, Weissman C. Kemper M. Continual trending 
of Fick variables in the critically ill patient. Chest 1991: 

188. Davies GG, Jebson PJR, Glascow BM. Hess DR. Con- 
tinuous Fick cardiac output compared to thermodilution 
cardiac output. Crit Care Med 1986;14:881-885. 

189. Davies G. Hess D. Jebson P. Continuous Fick cardiac 
output compared to continuous pulmonary artery elec- 
tromagnetic flow measurement in pigs. Anesthesiology 

190. Keinanen O. Takala J, Kari A. Continuous measurement 
of cardiac output by the Fick principle: clinical valida- 
tion in intensive care. Crit Care Med 1992;20:360-365. 

191. Mahutte CK, Jaffe MB. Sassoon CS, Wong DH. Cardiac 
output from carbon dioxide production and arterial and 
venous oximetry. Crit Care Med 1 99 1 : 1 9: 1 270- 1 277. 

192. Silverman HJ. Gastric tonometry: an index of splanch- 
nic tissue oxygenation? (editorial comment). Crit Care 
Med 1991;19:1223-1224. 

193. Gys T, Hubens A, Neels H, Lauwers LF. Peeters R. 
Prognostic value of gastric intramucosal pH in surgical 
intensive care patients. Crit Care Med 1988:16:1222- 

194. Landow L, Phillips DA, Heard SO, Prevost D. Vander- 
salm TJ, Fink MP. Gastric tonometry and venous oxim- 
etry in cardiac surgery patients. Crit Care Med 1991:19: 

195. Hartmann M, Montgomery A, Jonsson K. Haglund U. 
Tissue oxygenation in hemorrhagic shock measured as 
transcutaneous oxygen tension, subcutaneous oxygen 
tension, and gastrointestinal intramucosal pH in pigs. 
Crit Care Med 1991;19:205-210. 

196. Silverman HJ. Tuma P. Gastric tonometry in patients 
with sepsis: effects of dobutamine infusions and packed 
red blood cell transfusions. Crit Care Med 1992; 102: 




197. Gutierrez G. Bismar H, Dantzker DR. Silva N. Compari- 
son of gastric intramucosal pH with measures of oxygen 
transport and consumption in critically ill patients. Crit 
Care Med 1992:20:451-457. 

198. Gutierrez G, Palizas F, Doglio G, Wainsztein N. Gal- 
lesio A, Pacin J, et al. Gastric intramucosal pH as a ther- 
apeutic index of tissue oxygenation in critically ill pa- 
tients. Lancet 1992:339:195-199. 

199. Bentt LR. Santora TA. Leverle BJ. LoBue M. Shabot 
MM. Accuracy and utility of pulse oximetry in the sur- 
gical intensive care unit. Curr Surg 1990:47:267-268. 

200. Kestin IG, Miller BR. Lockhart CH. Auditory alarms 
during anesthesia monitoring. Anesthesiology 1988:69: 

201. Orlando R III. Continuous mixed venous oximetry in 
critically ill surgical patients: "high-tech' cost effective- 
ness. Arch Surg 1 986; 1 2 1 :470-47 1 . 

202. Pearson KS, Gomez MN, Moyers JR. Carter JG. Tinker 
JH. A cost/benefit analysis of randomized invasive mon- 
itoring for patients undergoing cardiac surgery. Anesth 
Analg 1989;69:336-341. 

203. Jastremski MS. Chelluri L. Beney KM. Bailly RT. Anal- 
ysis of the effects of continuous on-line monitoring of 
mixed venous oxygen saturation on patient outcome and 
cost effectiveness. Crit Care Med 1990; 1 7: 148- 1 53. 

204. Hein P. Grooks VI. Boorgen's pocketbooks 154. New 
York: Doubleday. 1978:30-31. 

Hess Discussion 

Dantzker: I had a fellow from a 
company producing intravascular 
blood gas monitors in my office the 
other day. My question to him was, 
Why do I need it? I reminded him of 
some interesting studies that were 
done a number of years ago in pa- 
tients during the onset of exercise. 1 
If you make measurements as rapid- 
ly as you can. by just drawing bloods 
during the onset of exercise, what 
you find in both the arterial and the 
mixed venous blood is a dramatic 
fall in Po: during that first 2 minutes 
or so of a ramp increase in exercise. 
The Po: then gradually comes back 
up toward normal, and this probably 
represents the system re-equilibrat- 
ing slowly, as oxygen is utilized at 
that initial stage. 

1. Edwards RHT, Denison DM, Jones 
G, Davies CTM. Campbell EJM. 
Changes in mixed venous gas ten- 
sions at start of exercise in man. J 
Appl Physiol 1972:32:165-169. 

Hess: The technical system, not the 
physiologic system? 

Dantzker: No, the physiologic sys- 
tem. As you begin to exercise and 
start using oxygen, the heart and the 
lungs take time to catch up. But they 
do catch up. 

Wood: Mixed venous Po : ? 
Dantzker: Both of them fall. Both 

arterial and mixed venous Po?. And I 
think back to those original studies 
of the mixed venous saturation cath- 
eter, in which they showed the won- 
derful pictures of a patient being 
turned and suctioned, and the drop- 
ping of his oxygen as reflected by 
the mixed venous Po2- I have always 
thought that that was good because it 
suggests that the tissues were ab- 
sorbing more oxygen to compensate 
for an increased metabolic need or a 
slight decrease in the arterial gas. I 
asked the salesman the same thing. I 
said, "How do we know what these 
minute-to-minute changes mean? 
Isn't it dangerous to start putting 
these things into the ICU without 
first understanding what this var- 
iability is, whether it's physiologi- 
cally important; how do you inter- 
cede on that basis? Wouldn't it be 
better if you just gave a whole bunch 
of good people a bunch of these de- 
vices to play with for a while, to see 
whether or not it was of any use?" 
And his answer to me was, "Well, 
you know, the company only has so 
many, and there's such a demand for 
them. They really can't afford to 
give you any because they're busy 
selling them to people who want to 
get them quickly into their ICUs." 

Hess: I guess my response would be 
that I hope we don't make the same 
mistake with this technology that we 
made with pulse oximetry. By the 
time we thought about whether it 

was good or bad, it was at the bed- 
side of every patient. 

East: I've looked at continuous in- 
vivo blood gas monitoring quite a bit. 
It's not quite as gloomy as you por- 
tray it. at the present time. Some of 
the data on existing systems really do 
look promising. The accuracy and 
precision look very nice. The ques- 
tion to list as Dr Dantzker put it. Are 
these things really useful? I mean, do 
we really need minute-to-minute 
blood gas monitoring? However. I 
think the technology is pretty good. 
Some of the new electrodes have the 
sensing material built into the clad- 
ding, so some manufacturing prob- 
lems that were inherent in earlier 
electrodes have been solved. 

Hess: I was just going to say that the 
preliminary data, at least the man- 
ufacturer's data, look pretty good. 
What I would like to see is more of 
that data in the peer-reviewed litera- 
ture, and. to my knowledge, there 
isn't much. 

East: Well, there is some.' : 

1. Larson CP. Siever A. Vendor J. Eval- 
uation of drift in an intra-arterial 
blood gas sensor (abstract). Anes- 
thesiology 1992;77(3A):A538. 

2. Scuderi PE. Bowton DL. Meredith 
JW. Harris LC. Evans JB. Anderson 
RL. A comparison of three pulmonary 
artery oximetry catheters in intensive 




care unit patients. Chest 1992,102: 

Hudson: Representatives of two 
manufacturers visited us recently 
(October 1992), and when we asked 
them for data, all they'd tell us about 
were data from the OR. You may 
have data from the ICU, but the rep- 
resentatives who were trying to sell 
these very expensive devices seem to 
have no data from the ICU. They 
know nothing about the effect of 
low-flow states. I really think we 
need to be tougher and not get into a 
buying frenzy just because a device 
is available (even if it looks good in 
relatively stable patients) until we 
know that it's technically good. And 
that doesn't even get to the question 
that Dave (Dantzker) brought up — 
How do we use it once we know that 
it really works? 

Hess: I think the other issue that you 
brought up is a very good one — just 
because it works well in the OR 
doesn't mean it's going to work in 
the ICU. 

Morris: If it does work in the ICU, 
once again we have to ask ourselves 
for what purpose, and then in what 
subset of patients — because it very 
likely is not going to have the same 
operational characteristics for every- 
body who might be in ICUs. I would 
caution us to be conservative about 
interpreting correlation coefficients, 
such as you presented in that table 
(Table 6, Page 659). The information 
in that table actually doesn't help 
anybody decide whether the device 
is good or bad, independent of the 
purpose for which the measurement 
is made. The correlation coefficient 
is of interest. It may be a necessary 
but insufficient descriptor of per- 

Hess: I agree 100%, and one of the 
points I made in the paper is that one 
limitation of all those studies is that 
they did not do the bias and the pre- 

cision type of analysis. All they 
reported was the correlation co- 
efficient. The study in Chest that 
came out in September 1992 does 
look at bias and precision. 1 The 
bias in that study was pretty good 
for all 3 devices, but the pre- 
cision, the limits of agreement, 
was really pretty wide. 

1. Scuderi PE, Bowton DL, Meredith 
JW. Harris LC, Evans JB. Anderson 
RL. A comparison of three pulmonary 
artery oximetry catheters in intensive 
care unit patients. Chest 1992;102: 

Morris: You helped us understand 
that for many of these devices, when 
the newer technology or the nonin- 
vasive technology was compared 
with what we would intuitively call 
the gold standard — for example, the 
CO-oximeter . . . 

Hess: Gold being very soft. 

Morris: Yeah. They were different. I 
suggest that we remember that — 
because for any one of these tech- 
nologies we have no idea how useful 
it is in actually directing clinical de- 
cision making. Even if there are dif- 
ferences between technologies, it is 
appropriate to remember that it may 
be that the noninvasive technique, 
though it may be less precise and 
even incorporate inaccuracies in the 
form of bias and random noise, 
works just as well as any other in 
guiding therapy. The problem is that 
we have relatively little information 
about how the information is used to 
foster clinical decision making. 

Hess: Good point. 

Wood: I think words matter, some- 
times. And monitoring means some- 
thing bad happens, the alarm goes 
off, and some team member responds 
to it. Ventricular fibrillation is a 
good example. It's really a good idea 
that the alarm goes off because re- 
sponding to it quickly saves lives. I 

don't think that mixed venous O; sat- 
uration ought to be dignified with the 
name "monitor," at least in the same 
way. If we didn't call it a monitor, as 
if there were an alarm to go off, a 
machine that told us something about 
patient care, we wouldn't expect it to 
do that — we wouldn't ask thera- 
peutic goals of it. Instead, we would 
use it as a measurement instrument 
that provides an answer to a question 
we ask. 

The continuous monitoring using 
dual oximetry of arterial and venous 
O: content would be a way to si- 
multaneously check whether an inter- 
vention that increased the cardiac 
output had the expected effect on the 
arterial-to-venous-0:-content differ- 
ence. One can then ask whether this 
is clinically relevant or not. but the 
expectation of the instrument is 
clear. It can either confirm your an- 
ticipated result or not. In those cir- 
cumstances, then, accuracy and pre- 
cision are two different expectations. 
I don't really need accuracy in order 
to follow the trends induced by an 
acute intervention as part of my ti- 
trated care of a critically ill patient to 
find out whether a variable has gone 
in the right direction. For example, 
when you put a patient on PEEP, can 
you use continuous O: saturation 
monitoring to get to the therapeutic 
goal — the least PEEP providing 90% 
saturation of an adequate circulating 
hemoglobin on a nontoxic Fio: — and 
get there very quickly? What some 
people can take an hour and a half to 
do can be done at the bedside in 5 
minutes, aided by continuous meas- 
urement of arterial saturation and 
blood pressure. That's following the 
direction of a change that tests the 
question. Can PEEP reduce the intra- 
pulmonary shunt and so increase ar- 
terial saturation without decreasing 
cardiac output and blood pressure? 
The problem with the term monitor 
is that it implies that there's an alarm 
somewhere other than the intensi- 
vist's intellect. And that's not true. 




There isn't one. I don't deny you can 
set the mixed venous saturation at 
90% and an alarm will go off, but 
that's much less information than 
you want. What you want is to inter- 
vene and see whether it achieves a 
therapeutic goal, and I think that's 
where the word monitoring gets us in 
trouble regarding device impact on 
critical care. 

Hess: I think part of what you're say- 
ing is that the issue becomes one of 
having the appropriate therapeutic 

Stoller: I want to make an epidemi- 
ologic comment with regard to ef- 
ficacy versus effectiveness. Most of 
the focus here is on the efficacy of 
the device, but, as you've mentioned, 
it needs to be put in the context of its 
clinical use. I think a quote from 
Thoreau is apt. He said, "The truth 
takes two — one to say it and one to 
hear it." I think, with regard to pulse 
oximetry, we're reminded of some 
data from Bowton and colleagues' in 
which continuous oximetry was used 
on the ward but frequently ignored. 
So, we have two imperatives, one to 
demonstrate the technical accuracy 
of the device, and the other to use the 
information intelligently. 

1. Bowton DL, Scuderi PE, Harris L. 
Haponik EF. Pulse oximetry mon- 
itoring outside the intensive care unit: 
progress or problem? Ann Intern Med 

Hudson: Maybe it's good to ignore 

East: I want to make one last point 
on monitoring devices, and it's a 
frustration we've had. We have ap- 
proached device manufacturers and 
asked them if they wanted us to do a 
study looking at efficacy. Of course, 
their response was that they don't 
have money for that kind of study; 
they don't need to do that to sell it. I 
think we, as a community, need to in- 
sist that devices have demonstrated 
efficacy. That'll force the community 
and the companies to put the money 
into doing the studies. It'll also put 
pressure on the federal funding agen- 
cies. You can't get money from the 
NIH to investigate efficacy of a de- 
vice; it just isn't there. And I think 
that's wrong. 

Pierson: By analogy, you cannot 
take your new cancer chemotherapy 
agent to the bedside and use it until it 
has been shown to be efficacious to 
the satisfaction of the FDA. And it's 
too bad that requirements aren't as 
stringent with some of these devices. 

Cerra: The new federal device laws 
are going into effect, which is going 
to correct some of this. There are 
standards of efficacy testing. Now, 
exactly how they're going to be ap- 
plied, I don't know. Certainly the 
new rules are not going to help any- 

thing that's already on the market or 
a look-alike that comes out on the 
market. So. the oximetry sorts of 
problems are here to stay. 

Hudson: I think the pulse oximeter 
is a very useful device, but when we 
start to monitor things or measure 
things continuously, we're going to 
find changes that we didn't know 
were there. They may not be bad or 
dangerous to the patient. It drives me 
crazy when every time the saturation 
falls transiently, the Fio: gets turned 
up, and then it takes a period of time 
to get it back down. We are reacting 
to changes that have happened all 
the time. We just didn't know about 
them, and they, presumably, aren't 
harmful. Now. obviously, if satura- 
tion falls and stays down for a period 
of time, then that's something we 
have to deal with. I think these are 
issues that we somehow need to ad- 
dress — how do we react to these 
changes, and what's clinically im- 
portant and what's not? 

Pierson: Alan Morris mentioned a 
few minutes ago something that I 
think should be a thread in this con- 
ference, and it's by no accident that 
we assigned him the very last talk — 
that is. getting back to the ultimate 
outcome of the patient benefit and its 
relation to some of the costs of hav- 
ing to do these things. 



The Nuts and Bolts of Increasing Arterial 
Oxygenation: Devices and Techniques 

Richard D Branson RRT 



Oxygen Therapy 

A. Fixed- vs Variable-Performance Equipment 

B. Systems & Devices for Oxygen Delivery 
C Selecting an Oxygen Delivery System 
End-Expiratory Pressure 

A. Design & Function of Valves 

B. Examples of EEP Valves 

C Clinical Importance of Valve & System Design 
D. System-Patient Interface 
Mean Airway Pressure Manipulation during 
Mechanical Ventilation 

A. Volume-Controlled Ventilation 

B. Pressure-Controlled Ventilation 
In Summary 


Providing therapy to reverse arterial hypoxemia 
involves the selection and appropriate application 
of one or more of a number of devices. Clinicians 
at the bedside should be familiar with the opera- 
tion, indications, complications, and limitations of 

The devices and techniques used to improve ar- 
terial oxygenation — delivery of supplemental oxy- 

Mr Branson is Assistant Professor, Division of Trauma/Critical 
Care, University of Cincinnati College of Medicine. Cin- 
cinnati. Ohio. 

A version of this paper was presented by Mr Branson on Oc- 
tober 8. 1992, during the RESPIRATORY CARE Journal Confer- 
ence on Oxygenation in the Critically 111 Patient, held in Puerto 
Vallarta. Mexico. 

The author has no financial interest in any of the products 
mentioned or in competing products. 

Reprints: Richard D Branson RRT. Division of Trauma/ 
Critical Care, University of Cincinnati College of Medicine, 
231 Bethesda Ave, Cincinnati OH 45267-0558. 

gen, increased alveolar ventilation (which I do not 
address), end-expiratory pressure (EEP), and ma- 
nipulation of mean airway pressure (P aw ) during 
mechanical ventilation — attempt to overcome the 
physiologic abnormalities that cause hypoxemia. 

Oxygen Therapy 

Fixed- vs Variable-Performance 

Oxygen therapy devices are typically classified 
as variable-performance (ie, low-flow) or fixed- 
performance equipment (ie, high-flow). Oxygen 
therapy techniques in the hospital have been well 
established for many years, and few changes have 
occurred in the past decade. In this review, I dis- 
cuss oxygen therapy devices in the manner used in 
many excellent textbooks' ~ 3 and report the results 
of recent investigations detailing function and ac- 

Low-flow, or variable-performance, equipment 
delivers oxygen at a fixed flow that provides only a 
portion of the patient's inspired gas. The term var- 
iable-performance relates to the fact that delivered 




oxygen is diluted with room air, and, thus, as the 
patient's ventilatory pattern changes, the fraction of 
inspired oxygen concentration (F102) fluctuates and 
may vary widely. In fact, despite some commonly 
published figures for delivered oxygen concentra- 
tion (Fdo:) at given flows, the concentration actual- 
ly received by the patient (ie, the F102) is neither 
precise nor predictable. Shapiro et al 3 have devised 
a method to predict F102 provided by variable- 
performance equipment using a number of assump- 
tions and a simplified equation. The equation as- 
sumes that (1) the anatomic reservoir is 50 mL, (2) 
the anatomic reservoir is filled with 100% oxygen 
prior to inspiration, and (3) ventilatory pattern (ti- 
dal volume, Vt: frequency, f; and inspiratory-to- 
expiratory-time ratio, I:E) is constant. 3 Table 1 
demonstrates the use of this equation assuming a 
V T of 500 mL, I:E of 1:2, f of 20 breaths/min, and 
inspiratory time of 1.0 second. Perhaps the most 
important lesson to be learned from this equation is 
the wide range of F102S possible with a constant 
flow of oxygen. 

Variable-performance equipment should not be 
routinely used in critical care because the patient's 
inspiratory flow may vary widely and such devices 
cannot ensure precise Fio2- Nasal cannulas are suit- 
ed for long-term use in stable patients requiring 
minimal oxygen supplementation. The simple mask 
offers a slightly higher F102 than the cannula but 
may be tolerated only for short periods of time. The 
successful use of the simple mask in the recovery 
room has been described. 4 The partial and non- 
rebreathing masks are acceptable for short-term use 
in emergency situations until definitive treatment 
can be provided. 

Fixed-performance equipment (including air-en- 
trainment masks, large-volume-nebulizer systems, 
and large-volume-humidifier systems) provides all 
of the patient's inspired gas at a precisely con- 
trolled Fdo2 and at a high flowrate. When the de- 
vice is applied correctly, the patient's F102 is con- 
stant, regardless of ventilatory pattern. 

Systems and Devices for 
Oxygen Delivery 

Nasal cannula. The nasal cannula is probably the 
most commonly used low-flow oxygen delivery de- 
vice. In theory, the nasal cannula increases F102 by 

Table 1 . Estimation of F102 from Low-Flow Systems* 

Cannula — 6 L/min 
Mechanical reservoir — 

Anatomic reservoir — 

50 mL 

100 mL 

Volume of (> inspired 
50 mL anatomic 

70 mL from room air 
220 mL O: inspired 

V T — 500 mL 
I:E— 1:2 

f— 20/min 
time — I s 

Because 150 mL of 
100% <> is inspired, 
the remainder of 
Vt is room air 
(350 mL), 20% of 
350 mL = 70 mL. 
amount of O2 in 
in room air inspired. 

FlO: = 

22(1 mL (O ) 
220mL(V T ) 

If Vt is decreased to 250 mL. 
volume O; inspired is 
50 mL anatomic reservoir 
100 mL flow/s 

20 mL from room air (20% of 250- 1 50 ) 
170 mL O: inspired 

170 mHO;) 
250mL(V T ) 


If Vt increased to 1 .000 mL 
volume Ch inspired is 
50 mL anatomic reservoir 
100 mL flow/s 

170 mL room air (20% of 1,000-150) 
320 mL O: inspired 

320mL(O : ) 
1,000 ixiL(Vt) 

= 0.32. 

* Adapted from Reference 2. with permission. 

0.04 for every liter of oxygen flow. Flow is typ- 
ically set at 2-6 L/min; flow above 6 L/min adds lit- 
tle to increased F102 and may produce patient dis- 
comfort, including nasal dryness and bleeding. The 
cannula is generally considered comfortable and 
well tolerated by the patient. 

Several authors have attempted to quantify the 
F102 afforded by the nasal cannula. Gibson and oth- 
ers 5 measured tracheal oxygen concentrations in 
two subjects via a transtracheal sampling catheter. 




Both subjects had tracheal oxygen concentration 
measured with a mass spectrometer during "quiet" 
breathing (f, 16 breaths/min; Vr, 400 mL; minute 
volume, Ve, 6.42 L/min); during "normal" breath- 
ing (f, 17 breaths/min; Vr, 690 mL; V E , 11 L/min); 
and during "hyperventilation" (f, 14 breaths/min; 
Vr, 1,400 mL; Ve, 19.5 L/min). Using a nasal can- 
nula at 1, 2, 3, 5, 10, and 15 L/min, they reported 
little variation in F102 at low inspiratory flows but 
wide variations at high inspiratory flows (Table 2). 
Schachter et al 5 studied oxygen delivery to ICU pa- 
tients. They found concentrations similar to those 
obtained by Gibson et al (Table 2). More recently, 
Ooi et al, 7 using a model of the respiratory system, 
studied Ficb afforded by the nasal cannula. Their 
model consisted of a face shield from a re- 
suscitation mannequin attached to a rubber test 
lung via tubing that approximated tracheal volume. 
The test lung was placed inside a rigid container, 
creating a bag-in-the-box system. A ventilator with 
a sine wave flow pattern was connected to the con- 
tainer, and inspiratory flow was varied from 12 to 
40 L/min at a constant f and Vr. Ooi et al found 
that Fio: varied from 13 to 40% at a given oxygen 
flow when the model's 'inspiratory flow' was var- 
ied from 12 to 40 L/min (Table 2). Dunlevy and 
Tyl 8 observed significantly higher Fio : s in 15 
healthy subjects during mouth-closed versus 
mouth-open breathing, with oxygen flow of 2 L/ 
min by nasal cannula. 

Table 2. Oxygen Concentrations Delivered by the Nasal Can- 


Measured Fdoi 







et aP 

etal 6 

et al 7 














































The nasal cannula is typically used to deliver ox- 
ygen to stable patients postoperatively and to pa- 
tients who require long-term therapy. Its use to 
combat hypoxemia in critically ill patients is prob- 

ably unwarranted because of the uncontrollable and 
widely variable Fk>. The use of bubble humidifiers 
with the nasal cannula is unnecessary as the in- 
crease in humidity is negligible. 910 

Simple mask. The simple mask can provide an in- 
crease in Fio: by extending the anatomic reservoir 
to include the volume within the mask. Simple 
masks are operated at flows of 5-12 L/min and are 
thus capable theoretically of providing Fio of 

Gibson et al 5 measured Ficb of 0.82-0.88 at a 
flow of 15 L/min. Redding et al" studied normal 
volunteers using a variety of oxygen delivery de- 
vices. This study deserves particular attention be- 
cause of the method used to determine Ficb. The 
authors solicited five normal volunteers from the 
respiratory therapy department to participate in the 
study. Each had a radial artery line placed to mon- 
itor P a o2- Oxygen was then delivered to the vol- 
unteers at precise oxygen concentrations from a 
high-flow blender and tight-fitting mask. Arterial 
blood was drawn, and P a cb was plotted against Fio: 
for each volunteer. The authors used regression 
analysis by the method of least squares to fit and 
plot a line. Arterial samples were then drawn while 
volunteers breathed from five types of oxygen 
mask. The P a o2 values obtained were then used to 
determine F102 by placement of datapoints on the 
fitted lines. Using this unique approach. Redding et 
al" found that at a flow of 6 L/min, F102 ranged 
from 0.38 to 0.46. No controls were used or meas- 
urements of breathing depth or frequency made. 
Bethune and Collis 12,13 studied the F102 received 
with simple masks and were among the first to sug- 
gest that a minimum flow was necessary to prevent 
rebreathing of carbon dioxide (CO : ). They found 
that at flows of 1 -8 L/min and a Vr of 500 mL, F102 
ranged from 0.21 to 0.60, and at a V T of 1,000 mL 
F102 ranged from 0.21 to 0.43. 

Milross et al 14 studied the Hudson Oxy-One face 
mask under laboratory conditions. They had a sin- 
gle trained volunteer, placed in a body ple- 
thysmograph, breathe at Vts of 0.3, 0.6, and 1.2 L. 
The volunteer breathed through the plethysmo- 
graphic mouthpiece. On the opposite side of the 
mouthpiece, the face mask was fitted to a plaster 
model of the volunteer's face. Oxygen concentra- 
tion was measured between the mask and the vol- 




unteer. At a constant f of 15 breaths/min, Vt was 
varied as described, and the mask was placed 
"tightly" or "loosely" on the face. Table 3 depicts 
the study's results. Milross et al 14 concluded that 
F102 with a simple mask is reliable and predictable 
as long as the mask is fitted properly to the face. 
However, they point out that position of the mask 
is difficult to control and that changes in I:E may 
alter results. 

Table 3. Oxygen Concentration Delivered (Fdo:) via Oxy- 
One Face Mask* 

Loose-Fitting Maskt Tight-Fitting Mask 
Minute Ventilation (L) 


5 12 20 

5 12 





0.36 0.31 0.28 

0.47 0.31 



0.46 0.37 0.33 

0.60 0.41 



0.47 0.40 0.35 

0.72 0.50 



0.46 0.42 0.38 

0.77 0.59 


t 1 5 breaths/rr 


*Adapted from Reference 4. with permi 
tBreathing frequency was maintained a 


It is generally said that a minimum flow of 5 L/ 
min is necessary to prevent rebreathing of C0 2 . 
This has recently been confirmed by Jensen et al. 15 
They found in studying normal volunteers that at 
flows < 5 L/min, Ve increased to maintain a con- 
stant PaCO:- It is rewarding when conventional wis- 
dom is confirmed by scientific investigation. 

Although bubble humidifiers are inefficient, 910 
my impression is that they are customarily used 
with simple masks. Whether a more efficient hu- 
midifier should be used is a subject for study. In 
my experience, complications of simple masks in- 
clude irritation from the mask or strap, claus- 
trophobia, and occasionally eye irritation from gas 
escaping around the bridge of the nose. 

Partial rebreathing mask. The partial rebreathing 
mask consists of a simple mask and a 1-L reservoir 
bag. The term partial rebreathing refers to the fact 
that the first one third of expired gas enters the res- 
ervoir bag. This is gas from the anatomic reservoir; 
so it does not contain C0 2 . As the bag fills from the 
oxygen flow and the first third of expiration, the re- 
maining expired gas exits the exhalation ports of 
the mask. Flow to the mask should be set so that 
during peak inspiratory flow the reservoir is not de- 

flated to less than one half of its volume. At lower 
flows, C0 2 may accumulate in the reservoir bag. 
Typically, flow to a partial rebreathing mask is set 
at 8-15 L/min, and the mask probably provides an 
F102 of 0.60-0.80. 

Few studies have measured F102 with the partial 
rebreathing mask. Kory et al lh found that at flows 
of 6 to 10 L/min the partial rebreathing mask pro- 
vided Fio: from 0.35 to 0.60. u As with any low- 
flow system, F102 can be expected to change as 
ventilatory pattern and oxygen flow are altered. 

Nonrebreathing mask. The nonrebreathing mask 
is similar to the partial rebreathing mask but con- 
tains three one-way valves. The first one-way valve 
is placed between the mask and the reservoir bag, 
prevents any expired gas from entering the res- 
ervoir, and aids in filling the reservoir with 100% 
oxygen. The other two one-way valves are posi- 
tioned on the exhalation ports of the mask and pre- 
vent entrainment of room air during peak in- 
spiratory demand. However, entrainment of room 
air can still occur around the mask. Oxygen flow 
should be set so that the reservoir bag does not col- 
lapse during inspiration. This typically requires a 
minimum flow of 10 L/min, and sometimes flows 
greater than 15 L/min are necessary. It is generally 
accepted that F102 with a nonrebreathing mask 
ranges from 0.60 to 0.80. 

Redding et al" found F102 of 0.57-0.70 when ox- 
ygen flow was greater than the patient's Ve. Early 
reports describing F102 near 1 .0 with the use of a 
nonrebreathing mask should be viewed with some 
skepticism. They may refer to the Boothby, Love- 
lace, Bulbulian (BLB) masks, which were tight- 
fitting, nondisposable masks with spring-loaded ex- 
piratory valves that completely prevented entrain- 
ment of room air. 17 Today's disposable devices 
cannot achieve similar results. 

Air-entrainment masks (AEMs). Air-entrainment 
masks have been the most frequently studied oxy- 
gen therapy devices,"*" 31 and are sometimes mis- 
labeled as "venturi masks" because of a mistaken 
notion of the operating principles."* Campbell de- 
veloped the original air-entrainment mask in 
I960 23 — a large, bulky system that included a se- 
ries of small exhaust holes. Today's commercially 
available designs are based on Campbell's original 




Fig. 1. Diagram of an air-entrainment mask that utilizes 
variable jet-orifice sizes to change Fdc>2- As diameter of 
the jet increases, gas velocity decreases, causing Fdo2 
to fall. (Reprited from Reference 2, with permission.) 

An AEM consists of the mask, a jet nozzle, and 
entrainment ports (Fig. 1). Oxygen under pressure 
is delivered through the jet nozzle located just be- 
low the mask. As gas travels through the jet nozzle, 
its velocity increases dramatically. Upon exiting 
the nozzle, the gas at high velocity entrains or 
drags ambient air into the mask. This is not due to 
the Bernoulli, or Venturi, principle but to viscous 
shearing forces between the gas traveling through 
the nozzle and the stagnant ambient air. ls The de- 
livered fraction of oxygen (Fdo:) depends on the 
size of the nozzle, the size of the entrainment ports, 
and the oxygen flow through the nozzle. Com- 
mercially available systems use either inter- 
changeable jets, adjustable entrainment ports, or a 
combination of the two. Typically, an AEM has 6 
to 8 F102 settings with a minimum suggested oxy- 
gen flow for each setting (Table 4). At F102 < 0.35, 

Table 4. F10; Setting. Minimum Flow Requirements. Outputs, 
and Entrainment Ratios for an Air-Entrainment 



Minimum Suggested 





Flow (L/min) 

: :Air 

Flow (L/min) 









































the AEM functions as a fixed-performance system. 
However, at F102 > 0.35, total flow falls below 40 
L/min causing the system to be considered a var- 
iable-performance system. 

Bubble humidifiers provide very little additional 
humidification when used with AEMs because of 
their inherent inefficiency and because of the com- 
paratively small oxygen flow compared to the en- 
trained flow. Humidity can be increased by using a 
large-volume nebulizer with bland solution and an 
aerosol collar that surrounds the entrainment ports. 
The dry oxygen then entrains air containing aero- 
sol, increasing humidity delivered to the patient. 
With the addition of aerosol particles to room air, 
entrained volume falls and Fdo: increases slightly. 

The accuracy of AEMs has been studied by nu- 
merous investigators. Campbell and Minty 22 have 
criticized commercially available AEMs because 
they believe the mask volume to be inadequate. 
They pointed out that as patient demand for flow 
increases, room air may be drawn in around the 
mask. In this instance, masks with larger volumes 
act as reservoirs for blended gas. assuring a con- 
stant Fio:- Cohen et al 21 monitored oxygen con- 
centration inside four AEMs in a laboratory study. 
They found a small but measurable difference be- 
tween set and delivered oxygen concentration. 
These differences were increased by rainout from 
an aerosol entrainment system. 

Woolner and Larkin 24 studied the Hudson Multi- 
vent mask in the laboratory to determine its ability 
to deliver a precise oxygen concentration. They 
used a face model connected to a pneumo- 
tachograph and mouthpiece to measure F102 during 
quiet breathing, with varying peak inspiratory flow. 
They found that a group of five masks provided ac- 
curate F102 at the 0.24, 0.26, and 0.28 settings. At 
the 0.30 setting and above, accuracy diminished. 
As oxygen flow was increased, the F102 afforded at 
the 0.30 and 0.35 setting improved, but at the 0.50 
setting it averaged only 0.39. Woolner and Larkin 21 ' 
also demonstrated that as peak inspiratory flow in- 
creased to 200 L/min, F102 fell dramatically. They 
concluded that total gas flow must be 30% greater 
than peak inspiratory flow at settings less than 0.30 
to ensure the desired Fio2- Woolner and Larkin 
concluded that above the 0.30 setting, the AEM is a 
variable-performance device. Fracchia and Torda 30 
studied five different AEMs and found that those 




with larger volume masks and extension tubes be- 
tween the jet and mask ensured more reliable and 
consistent Fio:- Cox and Gillbe 2 " investigated the 
ability of five AEMs to provide reliable F102 under 
laboratory conditions. They calculated the mean 
difference between Focb and F102 at Vt of 0.25 L, 
0.5 L, and 0.75 L. In agreement with other in- 
vestigators, they found that large-volume masks en- 
sured a more accurate Fio:- They also suggested 
that the aviation-style masks typically used are dis- 
advantageous because gas is delivered to the side 
of the nose and mouth; whereas with the original 
design by Campbell, gas is delivered directly to the 
nose and mouth. They concluded that in order for 
an AEM to ensure accurate F102. mask volume 
must be a minimum of 300 mL. 

Hill et al 27 found that mask volume was im- 
portant only if patient demand exceeded total flow. 
They suggested that low-volume masks with high 
flows were preferable to high-volume masks with 
lower flows. Lyew and colleagues 20 have modified 
AEMs to entrain additional oxygen so that any de- 
sired F102 may be obtained. The clinical usefulness 
of this concept has not yet been reported. 

Air-entrainment masks ensure Foot to patients 
requiring less than 0.35. Patients with chronic lung 
disease who may hypoventilate when exposed to 
high Fdo: are candidates for AEM usage. The 
AEM is intended for patients with high or changing 
ventilatory demands; therefore, use probably 
should be limited to these instances because of its 
high cost compared to variable-performance equip- 

Large-volume aerosol and humidifier systems. 

Large-volume aerosol systems use air-entrainment 
nebulizers alone or in parallel to provide gas to 
face masks, face tents, T-pieces, tracheostomy col- 
lars, and head hoods. Nondisposable aerosol sys- 
tems usually offer Fdo: °f 0.40, 0.60, and 1.0, 
whereas disposable systems offer 6 to 8 settings 
from 0.28 to 1.0. These systems use a constant jet- 
nozzle size with a variable-size entrainment port to 
change oxygen concentration. 

Under laboratory conditions, Foust et al 32 stud- 
ied F102 afforded by jet nebulizers in parallel. Us- 
ing Vt from 0.2 to 0.9 L and frequencies from 20 to 
40 breaths/min, they measured Fio: received at a 
lung model from parallel nebulizers set to deliver 

concentrations of 0.60, 0.80. and 1.0. Their study 
results are shown in Table 5. Their results demon- 
strate that under conditions of high ventilatory de- 
mand, such systems become variable-performance 
devices. This is particularly evident when data 
from the 0.80 and 1.0 settings are compared. De- 
spite the high-concentration setting, F102 falls be- 
cause the overall decrease in delivered flow results 
in entrainment of room air around the mask. These 
authors recommend that when precise F102 is nec- 
essary, a high-flow humidifier system be used. 

Table 5. Simulated Distal Airway O: Per Cent Received via 
Face Mask from Two Nebuli/ers in Parallel* 










Setting (%) 

Setting (%) 

Setting (%) 

Setting (%) 





Vt (mL) 

Ot % at Airway 

































51-58-52 48-54-50 
ce 32, with permission. 


* Adapted 

from Referen 

Kuo and colleagues" compared large-volume 
nebulizer systems to humidifier systems in 30 pa- 
tients receiving oxygen therapy. They found that at 
similar Fdo:S, P a 02 fell when patients were on aero- 
sol therapy. They attributed the fall to the adverse 
effects of aerosol therapy on lung function. How- 
ever, these authors did not measure actual F102 and 
therefore may have overlooked the findings of 
Foust et al. 32 It is doubtful that the changes in P a o2 
in this study were due to aerosol deposition. Rath- 
er, it is likely that F102 was l ess with tne air- 
entrainment nebulizer than with the large-volume 
humidification system because of a lower total 

Large-volume humidifier systems are the pre- 
ferred method of assuring accurate F102S when in- 
spiratory flowrates are high. Figure 2 depicts the 
system described by Foust et al. 32 Gas can originate 
from a blender, air and oxygen flowmeters, or a 




venturi system such as the Downs flow generator. 
All these devices are capable of delivering gas in 
excess of 100 L/min. Gas is then directed through a 
heated humidifier such as those used for mechan- 
ical ventilation. Standard 22-mm-ID tubing con- 
nects the humidifier to the aerosol mask, and a res- 
ervoir is often placed in-line. The system depicted 
in Figure 2 is capable of assuring an Fio: of 1.0 un- 
der all laboratory conditions studied by Foust et 
al. 32 

Fig. 2. High-flow humidification system for delivery of pre- 
cise Fdo2 up to 1 .0. (Reprinted from Reference 32, with 

Selecting an Oxygen Delivery System 

Choice of an oxygen delivery system requires 
knowledge of the patient's condition and of the 
therapeutic goals. Kacmarek suggests that asking 
the following questions will aid in choosing the ap- 
propriate device. 2 

What is the required Fio 2 ? 

Are consistency and accuracy of Fdo 2 required? 

Is an artificial airway present? 

What are the humidification requirements? 

Is tolerance and compliance a problem? 

The first of these questions immediately strat- 
ifies the various devices. A severely hypoxemic pa- 
tient rarely tolerates a nasal cannula, and a patient 
with mild respiratory distress probably does not 
need a nonrebreathing mask. The cannula does rep- 
resent the most frequently used 2 delivery device 
and is particularly useful for long-term application. 
Consistency of Fio 2 is then the next consideration. 
While the nasal cannula is appropriate in the stable 

COPD patient, an AEM is probably a better choice 
during acute exacerbations. The AEM allows pre- 
cise delivery at low Fdo 2 (< 0.30) despite changes 
in ventilatory pattern. The presence of an artificial 
airway and more rigorous humidification require- 
ments seem to go hand in hand. A patient with a 
tracheostomy is going to require either a large- 
volume nebulizer or a large-volume humidifier sys- 
tem to maintain integrity of the airway. In selected 
cases a hygroscopic condenser humidifier (HCH) 
with an oxygen nipple can be used. However, these 
devices are low-flow systems similar to a nasal 
cannula and are subject to the same limitations. 
The use of dry oxygen from a blender or high-flow 
system combined with an HCH has to my knowl- 
edge never been reported. This idea warrants fur- 
ther investigation. Finding the method most accept- 
able and best tolerated by a patient with an intact 
airway may be more difficult. Many patients find 
breathing warm humidified air uncomfortable 
when the upper airway is intact. In this instance, a 
high-volume nebulizer system may be preferred. If 
secretion clearance is not an issue, the AEM is 
ideal for low Fio 2 (< 0.50). 

We typically choose the nasal cannula for long- 
term use and for all noncritical cases. Patients with 
known COPD and unstable ventilatory patterns are 
placed on an AEM. In the emergency room, the 
simple, partial-, and nonrebreathing masks are used 
depending on the desired Fio:- These are used 
short-term until the patient reaches the ICU and de- 
finitive care can begin. In the ICU, we use high- 
flow-aerosol humidification systems for patients 
with artificial airways and either an AEM or aero- 
sol system for unintubated patients. When 100% 
oxygen is required, we prefer a high-flow humid- 
ification system and tight-fitting mask. This is usu- 
ally accomplished with a Downs flow generator 
and a wick humidifier set to provide a temperature 
of 28°C. Tolerance might seem unimportant, but a 
patient who refuses to use an oxygen delivery de- 
vice receives no benefit regardless of how well the 
system works. 

End-Expiratory Pressure 

The use of positive end-expiratory pressure 
(PEEP) and continuous positive airway pressure 
(CPAP) to increase functional residual capacity 
(FRC), decrease intrapulmonary shunt, and reduce 




the work of breathing is well documented in the lit- 
erature. 34 " 17 The components of a system to deliver 
PEEP/CPAP include a continuous- or demand-flow 
gas source, the system-patient interface, and an 
end-expiratory pressure valve. 

Design and Function of Valves 

Several authors have written lengthy reviews on 
the technical aspects of end-expiratory pressure 
(EEP) valves. 38 " 40 In this review, I describe the 
types of end-expiratory pressure valves and the ef- 
fects of EEP valves on the work of breathing. 

EEP valves are usually characterized as thresh- 
old resistors or flow resistors. 41 " 43 A threshold re- 
sistor is an EEP valve that allows for a relatively 
constant level of pressure regardless of flow. This 
is accomplished by application of a constant force 
opposing the flow. When the set pressure level is 
achieved, the EEP valve closes, causing flow to 
cease. A flow resistor creates a variable pressure, 
depending upon the amount of flow passing 
through it. A pure flow resistor maintains pressure 
only while flow is passing through it. If flow ceas- 
es, pressure falls to atmospheric pressure. This de- 
scription is simple, but — practically speaking — 
there are no pure flow or threshold resistors. All 
threshold resistors possess some degree of flow re- 
sistance related to the diameter of the components. 

Under ideal circumstances the function of a 
threshold resistor EEP valve can be expressed by: 



where P = pressure, F = force (in newtons), and SA = 
surface area in m 2 . 

The greater the force applied across a known sur- 
face area (eg, by spring tension), the greater the 
pressure developed. However, because all thresh- 
old resistors exhibit a degree of flow resistance, the 
equation must be modified to account for flow re- 

Poci-LVR 1 

where R 

resistance (cm H:0 ■ 

L- 1 ) and V= flow 

An EEP valve with low flow-resistive properties 
provides a relatively constant pressure across a 
specified flow range. If the EEP valve has high 
flow-resistive properties (increased R), then the 
pressure fluctuates with flow until flow ceases. 

The function of a true flow resistor is governed 
by the equation 


That is, when flow is absent the pressure remains 
equal to atmospheric pressure. In this case, the ad- 
justable variable is resistance. As resistance is in- 
creased, pressure rises linearly with flow. Alterna- 
tively, resistance can be constant and the pressure 
level adjusted by changing flow. 

Examples of EEP Valves 

Threshold-resistor EEP valves are either gravity- 
dependent (ie, position-dependent) or gravity-inde- 
pendent (ie, position-independent). 

Gravity-dependent devices. The gravity-depen- 
dent devices include the underwater-column, wa- 
ter-weighted diaphragm, and weighted-ball de- 
vices. In each case, the force is applied by a weight 
that remains constant only in an upright position. 
The underwater column is perhaps the oldest but 
least frequently used type of EEP valve. Expired 
gases are directed through a tower or column im- 
mersed in water. The height of the water above the 
outlet that must be displaced creates the force and 
hence the EEP level. The EEP level is adjusted by 
altering the depth of the column or the height of the 
water. The flow-resistive properties of underwater- 
column devices are primarily the result of the inter- 
nal diameter of the connecting tubes and column. 

The water-weighted diaphragm, often referred to 
as the Emerson water column, is shown in Figure 
3. 44 This device uses the weight of a column of wa- 
ter resting upon a diaphragm to create EEP. Ex- 
piratory gases from the patient are directed to the 
opposite side of the diaphragm. When the pressure 
created by expiratory flow exceeds the force ap- 
plied by the water, the diaphragm is displaced up- 
ward, and gas exits the outlet. When EEP equals 
the pressure created by the water's weight, the di- 
aphragm seals, maintaining PEEP or CPAP. The 
EEP level is adjusted by changing the height of the 
water in the column. This device is considered to 




Fig. 3. PEEP device with water- 
weighted diaphragm. (Reprinted from 
Reference 44, with permission.) 

have some flow-resistive properties; 4142 however, a 
change in the design to include a plastic screen be- 
tween the diaphragm and outlet may have resulted 
in higher flow-resistive properties than those pre- 
viously reported. 

The weighted-ball threshold resistor utilizes the 
weight of a ball over a known surface area to create 
PEEP. This device is similar in principle to the wa- 
ter-weighted diaphragm device, and its flow-resis- 
tive properties are related to the internal diameter 
of the gas outlet. 

Gravity-independent devices. The spring-loaded 
EEP valves are probably the most popular of the 
gravity-independent designs. The force is applied 
by a spring (or springs) against a disk that rests 
against a seat. When expiratory pressure exceeds 
the spring tension, the disk is lifted off its seat, al- 
lowing gas to vent to atmosphere. PEEP level is ad- 
justed by increasing or decreasing spring tension. 
Flow-resistive properties of spring-loaded devices 
are related to the spring characteristics, cross- 
sectional area for gas flow, and presence of tur- 
bulent flow. A common problem with spring- 
loaded devices is the fact that as the disk moves up- 
ward, spring tension changes as the spring is com- 
pressed. The Vital Signs PEEP valves have allevi- 
ated this problem by using dual concentric springs 
that maintain a relatively constant force regardless 
of spring compression. These devices are also com- 
paratively large, which helps to reduce flow re- 
sistance. Banner et al 43 found that of all spring- 
loaded devices, the Vital Signs PEEP valve most 
closely resembles a true threshold resistor. Other 
spring-loaded EEP devices typically have moderate 
flow-resistive properties. 

Magnetic PEEP valves utilize magnetic attrac- 
tion to oppose expiratory flow and create PEEP. 
Such devices use a ferromagnetic metal disk held 
against a valve seat by the magnetic field of a mov- 
able magnet. In order to unseat the disk, expiratory 
gas pressure must overcome the magnetic attrac- 
tion. PEEP is manipulated by adjusting the distance 
between the magnet and metal disk. Flow-resistive 
properties of these devices are related to the cross- 
sectional area for gas flow. Additionally, in my ex- 
perience, the initial pressure to unseat the disk is 
normally higher than the set PEEP level. This re- 
sults in a characteristic pressure spike at the very 
beginning of expiration, as seen on the airway pres- 
sure graphic. Commercially available magnetic de- 
vices often have moderate flow-resistive properties 
due to narrow diameters and the presence of tur- 
bulent flow. 

Exhalation-valve diaphragms and balloons. 

First- and second-generation ventilators often use 
mushroom valves, or inflatable diaphragms, as ex- 
halation valves. These are modified so that PEEP is 
maintained by partial inflation of the valve. The 
pressure in the balloon or diaphragm opposes cir- 
cuit pressure. When circuit pressure exceeds bal- 
loon pressure, the balloon is lifted from its seat and 
expiratory gases exit the outlet. Because the bal- 
loon has a greater surface area than the seat, pres- 
sure in the balloon is actually less than circuit pres- 
sure. Flow-resistive properties of these EEP valves 
are related to the cross-sectional area for gas flow, 
presence of turbulent flow, and the speed at which 
the balloon deflates. Marini et al 41 and Banner et 
al 42 have both described excessive flow resistance 
of these devices. Much of the resistance is created 
by slow balloon deflation, which is in turn the re- 




suit of the balloon inflation technique. Pressure in 
the balloon can be maintained by use of an adjust- 
able pressure regulator, a needle valve, or needle 
valve and venturi system. In the latter, for balloon 
deflation to occur, gas must travel back through the 
throat of the venturi. This can result in these de- 
vices' performing much more like flow resistors 
than like threshold resistors. 

Electromechanical valves. The newest generation 
microprocessor ventilators typically use an internal 
exhalation valve to control PEEP/CPAP. The elec- 
tromechanical valves use feedback control from 
measurements of circuit pressure. Banner et al 4: 
found that within this group of EEP valves some 
devices performed similarly to threshold resistors, 
whereas others behaved more like flow resistors. 
The major component of flow resistance in these 
devices is cross-sectional area for gas flow. Banner 
et al 43 found that, due to flow resistance, the scis- 
sors valve on the Siemens Servo 900C often pro- 
duced PEEP 10-15 cm H 2 higher than set PEEP. 

Clinical Importance of Valve & System Design 

Threshold vs flow-resistive valves. Threshold re- 
sistors are the preferred type of EEP valve. 4143 
They allow for a constant PEEP/CPAP level 
regardless of changes in expiratory flow. Threshold 
resistors are also associated with a lower in- 
spiratory work of breathing. 41 43 Discussions on the 
role of EEP-valve resistance in the production of 
auto-PEEP, pulmonary barotrauma, and car- 
diovascular compromise frequently end in dis- 
agreement. It seems prudent, however, to use 
threshold resistors, particularly with continuous- 
flow systems. In a system providing 80-90 L/min 
of continuous flow, the addition of the patient's ex- 
piratory flow at 60-70 L/min can result in drastic 
pressure changes if a flow-resistive device is used. 
Whether EEP valves contribute to auto-PEEP by 
increasing expiratory resistance remains unproven. 

Demand-flow vs continuous-flow systems. Few 

topics in mechanical ventilation have received as 
much attention in recent years as has the work of 
breathing (WOB) during mechanical ventilation. 
The literature is packed with comparisons of con- 

tinuous-flow and demand-valve systems and their 
contributions to patient WOB. 4? ''" An entire confer- 
ence could be convened on this subject — with no 
want for material. Because this topic is important, 
but peripheral to this discussion, I address only the 

Stand-alone continuous-flow systems are fre- 
quently used to provide CPAP and for many years 
were superior to demand-flow systems in mechan- 
ical ventilators. 46 ' 53 Regardless of the system used, 
the goal is to minimize the WOB. 

Microprocessor control of ventilators has im- 
proved the function of demand-valve systems 
through improved monitoring and processing of 
pressure signals and improved response of the flow 
valve itself. 61 Previously, all demand-valve systems 
used pressure as the trigger variable. Recently, sev- 
eral ventilators have been modified to use flow to 
trigger inspiration." 1 Although differences in ven- 
tilator systems are evident, for the most part the 
newest generation ventilators perform equivalent- 
ly. 45 Flow-triggering is associated with a slightly 
lower WOB than pressure-triggering. 61 However, 
this difference is related more to what occurs in the 
post-trigger phase than to the triggering alone. The 
flow-triggering system of the Puritan-Bennett 
7200ae uses a pressure target of 1 .5 cm H 2 above 
the PEEP level, and the pressure-triggering system 
uses a pressure target of the set sensitivity minus 
PEEP. The difference, then, at a sensitivity of -2 
cm H 2 on PEEP of 5 cm H : 0, is 3.5 cm H : of 
"pressure support." 61 It is incumbent upon the bed- 
side clinician to assure that WOB during PEEP/ 
CPAP is minimized by the appropriate setting of 
sensitivity and proper ventilator operation. 

Stand-alone systems for PEEP/CPAP can be used 
when mandatory breaths are not required. The sys- 
tem can be 'homemade' or purchased. At this writ- 
ing (1993), three commercially available systems 
provide a continuous flow adequate to maintain 
CPAP. One, the Downs flow generator, has an ad- 
justable Fdo: f rom 0.32 to 1.0 at a continuous flow 
of 90-130 L/min. 62 As CPAP increases. Fdo: in- 
creases and flow decreases due to reduced entrain- 
ment through the venturi. A second system, the 
AMBU system is set at a fixed entrainment ratio and 
provides an Fdo: of 0.33 with a flow of 20 L/min. 
The AMBU system incorporates an elastic-loaded 
reservoir bag designed to minimize airway- 



pressure fluctuation. This type of system is popular 
in Europe and has been reviewed extensively by 
Hillman et al. 63 ' 64 

The third flow generator (Respironics Sleep- 
Easy) operates electrically and was designed for 
the application of nasal CPAP in the home. The 
SleepEasy delivers room air at a flow of greater 
than 100 L/min. Supplemental oxygen can be bled 
into the system to provide an Fdo: ^ 0.35. As with 
any bleed-in system, the stability of the Fd(> is 
questionable. The advantages of this type of system 
are the savings of medical gas and the fact that a 
humidifier may not be required, depending on the 
humidity of the room air used. The clinical testing 
of this device for patients in respiratory failure has 
yet to be done. At present, in my experience, the 
Downs system is preferable for its high flowrate 
and adjustable Fio:- These systems all cost less 
than $300. 

System-Patient Interface 

PEEP/CPAP may be applied to an artificial air- 
way or via nasal or face mask. If masks are used, 
they should be lightweight and transparent and 
have an inflatable seal. 65 Transparent masks allow 
monitoring of the airway for the presence of secre- 
tions or vomitus. Application is aided by masks 
that are lightweight, and a soft, adjustable seal im- 
proves patient comfort. 

Mask CPAP is indicated for the treatment of hy- 
poxemia due to reduced FRC and ventilation/per- 
fusion abnormalities. Patients who are spontane- 
ously breathing, normo- or hypocarbic, with a P a oi/ 
Fio: > 150 and < 250, and who have the ability to 
protect their airway are candidates for mask 
CPAP. 65 Successful treatment has been demonstrat- 
ed in patients with pulmonary contusion, mild-to- 
moderate respiratory failure, pulmonary edema, 
COPD, asthma, and pneumonia. 65 

Mean Airway Pressure Manipulation during 
Mechanical Ventilation 

Mean airway pressure (P aw ), as a reflection of 
mean alveolar pressure (Pa), is one of the major de- 
terminants of oxygenation during mechanical ven- 
tilation. 66 7I Aside from the use of PEEP or CPAP, 

P aw may be altered by adjusting Vr, peak airway 
pressure, and the inspiratory flow or pressure wave- 
form. 70 

Volume-Controlled Ventilation (VCV) 

During VCV, the clinician may raise P aw by in- 
creasing Vr. During VCV, peak inspiratory pres- 
sure (PIP) is determined by the volume setting and 
patient resistance and compliance. Thus, Vr, PIP. 
and P uw are directly related and proportional. If Vj 
is increased sufficiently, depending upon in- 
spiratory flow, expiratory time may be insufficient 
to allow complete lung deflation. If this occurs, P aw 
will be further increased by the development of 

Increasing respiratory frequency at a constant Vr 
and inspiratory flow can increase P aw , predomi- 
nantly via the development of auto-PEEP, as ex- 
piratory time is shortened. If frequency is increased 
while inspiratory flow and time are held constant, 
rapid increases in P aw are seen. If the inspiratory 
time fraction (ti/t to t) is constant as frequency in- 
creases (ie, flow must increase), only small chang- 
es in P aw occur. This is true because P aw is the pres- 
sure-time component divided by total cycle time. If 
total cycle time decreases for a given PIP and in- 
spiratory time, P aw must increase. 

P avv can also be increased during VCV by de- 
creasing inspiratory flow or increasing ti/ttot- At a 
constant Vr and f, decreasing inspiratory flow re- 
sults in a long inspiratory time and a decreased to- 
tal cycle time and thus P aw is increased, by defini- 
tion. The relationship of flow and P aw at a given 
frequency is fairly constant until expiratory time is 
shortened enough to cause dynamic hyperinflation 
and auto-PEEP. After dynamic hyperinflation and 
auto-PEEP develop, small increases in ti/ttot can Je- 
suit in dramatic increases in auto-PEEP and P aw 
(Fig. 4). 

Further increases in P aw may be accomplished 
via an end-inspiratory pause. During an end- 
inspiratory pause, flow ceases and the end-inflation 
pressure maintains lung volume for the duration of 
the pause time. If the inspiratory pause causes a re- 
duction in expiratory time, auto-PEEP may de- 
velop, further increasing P aw . 

Changes in the inspiratory flow waveform and 
consequently the airway pressure waveform can 





Respiratory Cycle 

Fig. 4. Options for increasing mean airway and alveolar 
pressures during passive inflation with constant flow. The 
shaded areas represent alveolar pressure. For the same 
inspiratory time (ti) and Vt, both mean airway pressure 
and mean alveolar pressure can be increased by in- 
creasing the positive end-expiratory pressure (PEEP) or 
creating auto-PEEP, by lengthening the end-inspiratory 
pause, or by 'squaring-up' the inflation pressure profile 
(dashed lines). Lengthening the pause period may ac- 
centuate auto-PEEP if it shortens the deflation period 
(t e ). PEEP and plateau areas can be considered static 
pressures that together with the inspiratory tidal elastic 
pressure are shared in common by airway pressure and 
alveolar pressure. Pd = dynamic pressure; Ps = static 
pressure. (Reprinted from Reference 92, with permis- 

also alter P aw . For a given Vr and tj/t to t. a de- 
celerating flow waveform causes an increase in P aw 
because peak inflation pressure is reached earlier in 
the inspiratory period and maintained throughout 
the inspiratory time. 72 Rau and Shelledy found that 
for a given Vt, P aw was greatest with a decelerating 
flow waveform followed by sine, square, and ac- 
celerating flow waveforms. 72 A similar effect can 
be obtained during VCV by increasing inspiratory 
flow (allowing PIP to be reached earlier) and add- 
ing an inspiratory hold to maintain a constant tj/ttot. 
Figure 4 illustrates the methods for increasing P aw 
during VCV, as described by Marini and Ra- 
venscraft. 70 

The clinical effect of these manipulations has 
been described by several authors, with differing 
results." 87 No clinical studies suggest that any of 
these methods is superior for maintaining oxygena- 
tion or reducing morbidity or mortality. 

Pressure-Controlled Ventilation 

During the last 5 years, the technique of pres- 
sure-controlled ventilation has been resurrected. 

represented as being 'new,' and proclaimed to be a 
panacea for the treatment of respiratory failure. 88 " 91 
PCV allows for similar manipulations of P aw ; how- 
ever, relationships among variables change. During 
PCV, peak inflation pressure is preset and Vt is de- 
pendent upon pulmonary impedance, tj/ttot, and f. w: 
Marini and Ravenscraft have calculated P aw during 
PCV as 

d ..(Psc.Mt.l MPEEP)(te) 

where P se t is the set pressure level, t j = inspiratory time, 
ttot = total cycle time, and t e = expiratory time. 

Increasing P aw during PCV can be accomplished 
by increasing P se t (similar to increasing Vt in 
VCV). However, if auto-PEEP develops as Vt in- 
creases, Vt for a given P se t may actually fall. This 
occurs when the presence of auto-PEEP causes the 
absolute pressure change (PIP - total PEEP) to di- 

Increases in P aw may also be accomplished by 
increasing tj and shortening t e with a resultant fall 
in ttot- Changes in respiratory frequency at a given 
Pset and tj/ttot do not increase P aw . As with VCV, 
when t e is shortened sufficiently, auto-PEEP de- 
velops, causing P aw to increase. Inverse ratio ven- 
tilation (IRV) utilizes this phenomenon to increase 
P aw while decreasing PIP. During IRV. P aw in- 
creases with increases in P se t. tj, and auto-PEEP. If 
inspiratory and expiratory resistances are equivalent, 
P aw increases linearly with tj/t l0 t. 

The choice of ventilatory strategy is greatly in- 
fluenced by clinician experience, patient popula- 
tion, technical expertise, and available equipment. 
At present, no one method of increasing P aw is 
known to be superior, but appropriate monitoring 
and clinician vigilance are imperative. 

In Summary 

P a o: can be increased by oxygen supple- 
mentation or by application of EEP and by ma- 
nipulation of P aw during mechanical ventilation. 
Techniques and devices for accomplishing these 
maneuvers are many and varied. Clinicians should 
be aware of the appropriate application, complica- 
tions, and limitations of each. 





Ward J. Equipment for mixed gas and oxygen therapy. 
In: Barnes TJ, ed. Respiratory care practice. Chicago: 
Year Book Medical Publishers, 1988:287-355. 
Kacmarek RM. Methods of oxygen delivery in the hos- 
pital. In: Christopher KL. ed. The current status of oxy- 
gen therapy. Probl Respir Care 1990;3:563-574. 
Shapiro BA, Harrison RA. Kacmarek RM. Cane RD. 
Oxygen therapy. In: Shapiro BA. Harrison RA, Kac- 
marek RM. Cane RD, eds. Clinical application of res- 
piratory care, 3rd ed. Chicago: Year Book Medical Pub- 
lishers, 1985:176-191. 

Hudes ET, Marans HJ, Hirano GM, Scott AC, Ho K. 
Recovery room oxygenation: a comparison of nasal 
catheters and 40 percent oxygen masks. Can J Anaesth 

Gibson RL, Comer PB, Beckham RW, McGraw CP. 
Actual tracheal oxygen concentrations with commonly 
used oxygen equipment. Anesthesiology 1976:44:71-73. 
Schachter EN, Littner MR, Luddy P, Beck GJ. Mon- 
itoring of oxygen delivery systems in clinical practice. 
Crit Care Med 1980;8:405-409. 

Ooi R, Joshi P, Soni N. An evaluation of oxygen de- 
livery using nasal prongs. Anesthesia 1992:47:591-593. 
Dunlevy CL, Tyl SE. The effect of oral versus nasal 
breathing on oxygen concentrations received from nasal 
cannulas. Respir Care 1992;37:357-360. 
Estey W. Subjective effects of dry versus humidified 
low flow oxygen. Respir Care 1980;25:1143-1 144. 
Campbell EJ, Baker MD, Crites-Silver P. Subjective ef- 
fects of humidification of oxygen for delivery by nasal 
cannula: a prospective study. Chest 1988;93:289-293. 
Redding JS, McAfee DD, Gross CW. Oxygen con- 
centrations received from commonly used delivery sys- 
tems. South Med J 1978;71:169-172. 
Bethune DW, Collis JM. The evaluation of oxygen 
masks: a mechanical method. Anaesthesia 1967;22:43- 

Bethune DW, Collis JM. Evaluation of oxygen therapy 
equipment: experimental study of various devices on the 
human subject. Thorax 1967;22:221-225. 
Milross J, Young IH, Donnelly P. The oxygen delivery 
characteristics of the Hudson Oxy-one face mask. An- 
aesth Intensive Care 1 989; 1 7: 1 80- 1 84. 
Jensen AG. Johnson A, Sandstedt S. Rebreathing during 
oxygen treatment with face mask: the effect of oxygen 
flow rates on ventilation. Acta Anaesthesiol Scand 

Kory RC, Bergman JC. Sweet RD, Smith JR. Compar- 
ative evaluation of oxygen therapy techniques. JAMA 

Boothby WM. Lovelace WR. Vihlein A. The BLB oxy- 
gen inhalation apparatus: improvements in design and 
efficiency by studies on oxygen percentages in alveolar 
air. Proc Mayo Clin 1940;15:194-206. 

1 8. Scacci R. Air entrainment masks: jet mixing is how they 
work; the Bernoulli and Venturi principles are how they 
don't. Respir Care 1979;24:928-931. 

19. Friedman SA. Weber B. Briscoe WA. Smith JP. King 
TKC. Oxygen therapy: evaluation of various air-en- 
training masks. JAMA 1974;228:474-478. 

20. Lyew MA, Holland AJ, Metcalf IR. Combined air and 
oxygen entrainment: effect on the percentage output of 
fixed performance masks. Anaesthesia 1990:45:732- 

21. Cohen JL. Demers RR. Saklad M. Air-entrainment oxy- 
gen masks: a performance evaluation. Respir Care 1977; 

22. Campbell EJM. Minty KB. Controlled oxygen therapy 
at 60% concentration: why and how. Lancet 1976; 1 : 

23. Campbell EJM. A method of controlled oxygen ad- 
ministration which reduces the risk of carbon dioxide re- 
tention. Lancet 1960;2:12-14. 

24. Campbell EJM. Apparatus for oxygen administration. 
Br Med J 1963;2:1269-1270. 

25. Campbell EJM. Gebbie T. Masks and tent for providing 
controlled oxygen concentrations. Lancet 1966:1:468- 

26. Speir WA Jr. Kaplan HM. Weir MA, Ellison LT. Oxy- 
gen concentration delivered by Venturi masks with in- 
line humidification. JAMA 1971;216:879-880. 

27. Hill SL. Barnes PK, Hollway T, Tennant R. Fixed per- 
formance oxygen masks: an evaluation. Br Med J Clin 
Res Ed 1984;288:1261-1263. 

28. Cox D, Gillbe C. Fixed performance oxygen masks: hy- 
poxic hazard of low-capacity drugs. Anaesthesia 1981; 

29. Woolner DF, Larkin J. An analysis of the performance 
of a variable venturi-type oxygen mask. Anaesth Inten- 
sive Care 1980;8:44-51. 

30. Fracchia G, Torda TA. Performance of Venturi oxygen 
delivery devices. Anaesth Intensive Care 1980:8:426- 

31. Canet J. Sanchis J. Performance of a low-flow O: Ven- 
turi mask: diluting effects of the breathing pattern. Eur J 
Respir Dis 1984;65:68-73. 

32. Foust GN, Potter WA, Wilons MD, Golden EB. Short- 
comings of using two jet nebulizers in tandem with an 
aerosol face mask for optimal oxygen therapy. Chest 

33. Kuo CD, Lin SE, Wang JH. Aerosol, humidity and oxy- 
genation. Chest 1991;99:1325-1356. 

34. Ashbaugh DG, Bigelow DB. Petty TL. Levine BE. 
Acute respiratory distress in adults. Lancet 1967:2:319- 

35. Mclntyre RW. Laws AK. Ramachandran PR. Positive 
expiratory pressure plateau: improved gas exchange dur- 
ing mechanical ventilation. Can Anaesth Soc J 1969; 16: 

36. Gregory GA, Kitterman JA. Phibbs RH. Tooley WH. 
Hamilton WK. Treatment of the idiopathic respiratoiy- 




distress syndrome with continuous positive airway pres- 
sure. N Engl J Med 1971;284:1333-1340. 

37. Kirby RR, Downs JB, Civetta JM. Modell JH, Danne- 
miller FJ, Klein EF. et al. High level positive end ex- 
piratory pressure (PEEP) in acute respiratory insuffi- 
ciency. Chest 1975:67:156-163. 

38. Kacmarek RM. Dimas S, Reynolds J. Shapiro BA. 
Technical aspects of positive end-expiratory pressure 
(PEEP) Parts I, II. and III. Respir Care 1982;27:1478- 

39. Banner MJ. Expiratory positive-pressure valves: flow 
resistance and work of breathing. Respir Care 1987:32: 

40. Spearman CB. Positive end-expiratory pressure: ter- 
minology and technical aspects of PEEP devices and 
systems. Respir Care 1988;33:434-443. 

41. Marini JJ, Culver BH, Kirk W. Flow resistance of ex- 
halation valves and positive end-expiratory pressure de- 
vices used in mechanical ventilation. Am Rev Respir 
Dis 1985;131:850-854. 

42. Banner MJ, Downs JB, Kirby RR, Smith RA, Boysen 
PC Lampotang S. Effects of expiratory flow resistance 
on the inspiratory work of breathing. Chest 1988; 

43. Banner MJ, Lampotang S, Boysen PG, Hurd TE, De- 
sautels DA. Flow resistance of expiratory positive- 
pressure valve systems. Chest 1986;90:212-217. 

44. Spearman CB, Sanders HG. Physical principles and 
functional designs for ventilators. In: Kirby RR, Smith 
RA, Desautels DA, eds. Mechanical ventilation. New 
York: Churchill Livingstone. 1985:59-1 13. 

45. Hirsch C, Kacmarek RM, Stanek K. Work of breathing 
during CPAP and PSV imposed by the new generation 
mechanical ventilators: a lung model study. Respir Care 

46. Op't Holt TB, Hall MW, Bass JB, Allison RC. Com- 
parison of changes in airway pressure during continuous 
positive airway pressure (CPAP) between demand valve 
and continuous flow devices. Respir Care 1982:27: 

47. Viale JP, Annat G, Bertrand O, Godard J, Motin J. Ad- 
ditional inspiratory work in intubated patients breathing 
with continuous positive airway pressure systems. An- 
esthesiology 1985;63:536-539. 

48. Henry WC, West GA, Wilson RS. A comparison of the 
oxygen cost of breathing between a continuous-flow 
CPAP system and a demand-flow CPAP system. Respir 
Care 1983;28:1273-1281. 

49. Cox D, Niblett DJ. Studies on continuous positive air- 
way pressure breathing systems. Br J Anaesth 1984; 

50. Christopher KL, Neff TA, Bowman JL. Eberle DJ. Irvin 
CG, Good JT Jr. Demand and continuous flow inter- 
mittent mandatory ventilation systems. Chest 1985:87: 

51. Katz J A. Kraemer RW, Gjerde GE. Inspiratory work 
and airway pressure with continuous positive airway 
pressure delivery systems. Chest 1985;88:519-526. 

52. Capps JS, Ritz R. Pierson DJ. An evaluation, in four 
ventilators, of characteristics that affect work of breath- 
ing. Respir Care 1987:32:1017-1024. 

53. Gibney RTN. Wilson RS, Pontoppidan H. Comparison 
of work of breathing on high gas flow and demand 
valve continuous positive airway pressure systems. 
Chest 1982;82:692-695. 

54. Street MK, Hopkinson RB. Evaluation of the comfort of 
spontaneous respiration through three ventilator sys- 
tems. Intensive Care Med 1987;13:405-410. 

55. Beydon L, Chasse M. Harf A. Lemaire F. Inspiratory 
work of breathing during spontaneous ventilation using 
demand valves and continuous flow systems. Am Rev 
Respir Dis 1988:138:300-304. 

56. Samodelov LF. Falke KJ. Total inspiratory work with 
modern demand valve devices compared to continuous 
flow CPAP. Intensive Care Med 1988;14:632-639. 

57. Cox D, Tinloi SF. Farrimond JG. Investigation of the 
spontaneous modes of breathing of different ventilators. 
Intensive Care Med 1988;14:532-537. 

58. Bersten AD, Rutten AJ, Vedig AE. Showronski GA. 
Additional work of breathing imposed by endotracheal 
tubes, breathing circuits and intensive care ventilators. 
Crit Care Med 1989;17:671-677. 

59. Sassoon CSH, Giron AE, Ely EA, Light RW. Inspir- 
atory work of breathing on flow-by and demand-flow 
continuous positive airway pressure. Crit Care Med 

60. Saito S, Tokioka H, Kosaka F. Efficacy of flow-by dur- 
ing continuous positive airway pressure ventilation. Crit 
Care Med 1990;18:654-656. 

61. Sasson CSH. Mechanical ventilator design and function: 
the trigger variable. Resp Care 1992;37:1056-1069. 

62. Ward JJ, Awan ZI. Flow capabilities of the adjustable 
Downs flow generator (abstract). Respir Care 1987:32: 

63. Hillman DR. Finucane KE. Continuous positive airway 
pressure: a breathing system to minimize respiratory 
work. Crit Care Med 1985;13:38-43. 

64. Hillman DR. Breakey JN. Lam YM. Noffsinger WJ. 
Finucane KE. Minimizing work of breathing with con- 
tinuous positive airway pressure and intermittent man- 
datory ventilation: an improved continuous low-flow 
system. Crit Care Med 1987;15:665-670. 

65. Branson RD. PEEP without endotracheal intubation. Re- 
spir Care 1988;33:598-612. 

66. Boros SJ, Matalon SV, Ewald R, Leonard AS, Hunt CE. 
The effect of independent variations in inspiratory-ex- 
piratory ratio and end-expiratory pressure during me- 
chanical ventilation in hyaline membrane disease: the 
significance of mean airway pressure. J Pediatr 1977:91 : 

67. Ciszek TA. Modanlou HD, Owings D, Nelson P. Mean 
airway pressure — significance during mechanical ven- 
tilation in neonates. J Pediatr 1981:99:121-126. 

68. Stewart AR, Finer NN, Peters KL. Effects of alterations 
of inspiratory and expiratory pressures and inspiratory/ 













expiratory ratios on mean airway pressure, blood gases, 80. 

and intracranial pressure. Pediatrics 1981;67:474-481. 
Pesenti A, Marcolin R, Prato P, Borelli M, Riboni A, 
Gattinoni L. Mean airway pressure vs positive end-ex- 
piratory pressure during mechanical ventilation. Crit 81. 
Care Med 1985;13:34-37. 

Marini JJ. Ravenscraft SA. Mean airway pressure: phys- 
iologic determinants and clinical importance — Part 1 : 82. 
physiologic determinants and measurements. Crit Care 
Med 1992;20:1461-1472. 

Marini JJ. Ravenscraft SA. Mean airway pressure: phys- 
iologic determinants and clinical importance — Part 2: 83. 
clinical implications. Crit Care Med 1992:20:1604- 

Rau JL Jr, Shelledy DC. The effect of varying in- 
spiratory flow waveforms on peak and mean airway 84. 
pressures with a time-cycled volume ventilator: a bench 
study. RespirCare 1991;36:347-356. 
Bergman NA. Effect of varying respiratory waveforms 
on distribution of inspired gas during artificial ventila- 85. 
tion. Am Rev Respir Dis 1969;100:518-525. 
Adams AP, Economides AP, Finlay WEI, Sykes MK. 
The effects of variations of inspiratory flow waveform 86. 
on cardiorespiratory function during controlled ventila- 
tion in normo-, hypo-, and hypervolaemic dogs. Br J 
Anaesth 1970;42:818-825. 87. 
Baker AB, Cowie RW, Colliss JE. Effects of varying in- 
spiratory flow waveform and time in intermittent pos- 
itive pressure ventilation. III. Blockade of the auto- 
nomic nervous system. Br J Anaesth 1977;49:1235- 88. 

Branson RD, Hurst JM. Effects of inspiratory flow pat- 
tern on airway pressure, ventilation, and hemodynamics 
(abstract). RespirCare 1987;32:913. 89. 

Smith RA, Venus B. Cardiopulmonary effect of various 
inspiratory flow profiles during controlled mechanical 
ventilation in a porcine lung model. Crit Care Med 90. 


Baker AB, Babington PCB, Colliss JE, Cowie RW. Ef- 
fects of varying inspiratory flow waveform and time in 91. 
intermittent positive pressure ventilation. I: introduction 
and methods. Br J Anaesth 1977;49:1207-1220. 

Baker AB, Colliss JE, Cowie RW. Effects of varying in- 92. 

spiratory flow waveform and time in intermittent pos- 
itive pressure ventilation. II: various physiological var- 
iables. Br J Anaesth 1977;49:1221-1234. 

Al-Saady N, Bennett ED. Decelerating inspiratory flow 
waveform improves lung mechanics and gas exchange 
in patients on intermittent positive-pressure ventilation. 
Intensive Care Med 1985:11:68-75. 
Modell HI. Cheney FW. Effects of inspiratory flow pat- 
tern on gas exchange in normal and abnormal lungs. J 
Appl Physiol 1979;46:1103-1 107. 
Baker AB. Thompson JB, Turner J. Hansen P. Effects 
of varying inspiratory flow waveform and time in inter- 
mittent positive pressure ventilation: pulmonary oe- 
dema. Br J Anaesth 1982;54:539-546. 
Baker AB, Restall R, Clark BW. Effects of varying in- 
spiratory flow waveform and time in intermittent pos- 
itive pressure ventilation: emphysema. Br J Anaesth 

Banner MJ, Boysen PC Lampotang S. Jaeger MJ. End- 
tidal CO: affected by inspiratory time and flow wave- 
form — time for a change (abstract). Crit Care Med 

Lyager S. Influence of flow pattern on the distribution 
of respiratory air during intermittent positive-pressure 
ventilation. Acta Anaesthesiol Scand 1 968: 12:191-211. 
Jansson L, Jonson B. A theoretical study on flow pat- 
terns of ventilators. Scand J Respir Dis 1972:53:237- 

Johansson H, Lofstrom JB. Effects of breathing me- 
chanics and gas exchange of different inspiratory gas 
flow patterns during anaesthesia. Acta Anaesthesiol 
Scand 1975;19:8-18. 

Gurevitch MJ. Van Dyke J, Young ES, Jackson K. Im- 
proved oxygenation and lower peak airway pressure in 
severe adult respiratory distress syndrome: treatment 
with inverse ratio ventilation. Chest 1986:89:21 1-213. 
Tharratt RS, Allen RP, Albertson TE. Pressure con- 
trolled inverse ratio ventilation in severe adult res- 
piratory failure. Chest 1988;94:755-762. 
Cole AGH, Weller SF, Sykes MK. Inverse ratio ventila- 
tion compared with PEEP in adult respiratory failure. 
Intensive Care Med 1984;10:227-232. 
Abraham E, Yoshihara G. Cardiorespiratory effects of 
pressure-controlled inverse ratio ventilation in severe 
respiratory failure. Chest 1989;96:1356-1359. 
Marini JJ. Crooke PS III, Truwit JD. Determinants and 
limits of pressure-preset ventilation: a mathematical 
model of pressure control. J Appl Physiol 1989:67: 




Branson Discussion 

Dantzker: Rich, you told us a little 
bit about the consistency of F102 — or 
the inconsistency — with both types 
of delivery systems by mask. What 
about with ventilators? What's the 
consistency or inconsistency in Fio 2 
with mechanical ventilation? 

Branson: I have a lot of experience 
with this because one of the things 
that I do a lot is make metabolic 
measurements, and the stability of 
F102 is absolutely paramount there. In 
most ventilators, the Fdo2 and thus 
the Fio: during a breath changes as 
much as 2.5-3%. It also all depends 
on the ventilator setting. At 60%. 
where the ratio is 1:1, oxygen con- 
centration tends to be more constant 
than at 80% or at 30%. But the con- 
sistency is probably reasonable for 
the clinical care that we do — 2, 3, 
4% from one minute to the next — but 
not for metabolic measurements. 

Dantzker: So how do you make 
metabolic measurements? We've all 
been talking about measuring \fo 2 by 
expired gas. How would you do it on 
somebody on a ventilator. I remem- 
ber back in the 70s, when we were 
playing around with the first-gen- 
eration volume ventilators. I thought 
maybe they'd improved to the point 
where you can actually get reason- 
able . . . 

Branson: There are two things to do. 
The more expensive way is to use an 
air-oxygen blender. The biggest prob- 
lem is that your oxygen and air pres- 
sures in a hospital are never equal. 
When you plug the blender in, even 
if the blender works by differential 
pressure, it allows some gas to es- 
cape from one side to equalize the 
pressure, which changes your Fdo:- 
So what we do is put little regulators 
on the quick connects that go into the 
wall outlets to maintain the pressure 
going to the blender constant from 
both lines. Then we send gas to the 

blender, and the blender gives us a 
fairly reasonable Fdo: and Fio?. Then 
when gas goes into the ventilator and 
gets processed through the ven- 
tilator's blender, the ventilator is 
blending two gases of the same con- 
centration. Then you can actually get 
precise oxygen concentrations. That's 
very important. One of the things 
that affects F102 is the minute ventila- 
tion from ventilators — the higher the 
minute ventilation, the more unstable 
the oxygen concentration. The sec- 
ond thing that we do, which we re- 
cently presented at ASPEN, 1 is to use 
an inspiratory mixing chamber. We 
take a mixing chamber (about 3 liters 
in size) from a metabolic cart, and 
put it on the outlet of the ventilator, 
and it actually mixes the gas before it 
reaches the patient. At low Fio:S and 
minute ventilations (less than 10 L/ 
min), that's a less costly, less time- 
consuming procedure that stabilizes 
the F102 to a fairly constant level. 

1. Branson RD, Davis K, Johnson DJ. A 
simple method for stabilizing oxygen 
concentration during metabolic meas- 
urements (abstract). JPEN 1992;16: 

Dantzker: But you would think that 
unless you do something like that, 
the variability of F102 is going to be 
really dramatic. 

Branson: You cannot make meta- 
bolic measurements without doing 
one of those two things. Again, the 
ventilators make a difference, too. 
The Hamilton Veolar has an 8-L res- 
ervoir between the blender and the 
outlet valves, so it has a sort of in- 
spiratory mixing chamber. The Sie- 
mens 900C with a spring-loaded bel- 
lows tends to act as an inspiratory 
mixing chamber also, but the 7200 
ventilators that do on-line mixing 
from proportional solenoids, espe- 
cially during pressure-control venti- 
lation where you have this very rapid 
rise in pressure ... I agree. We heard 
the talk this morning about the F102 

and making measurements up to 
80% with a metabolic cart. We can 
all do that in the laboratory, especial- 
ly if we use tonometered gases in- 
stead of a blender. But when we take 
it out to the ICU, with an open- 
circuit system, we're lucky to be able 
to measure somebody on greater than 
60% oxygen. Consider yourself very 
lucky if you get good measurements 
at Fio 2 s > 0.60. 

Hess: Rich, you just touched on a 
point I was going to make, and that 
is that it really depends upon the type 
of system you use. If you use the 
open-circuit system, what you're 
talking about is a problem. But if you 
use the closed-circuit system, then 
F102 fluctuation becomes virtually a 
nonexistent problem. Isn't that true, 

Branson: The closed-circuit system 
uses volume loss of oxygen from the 
reservoir over time. The problem with 
that is that it increases the work of 
breathing during spontaneous breath- 
ing and doesn't reliably reproduce 
the mode of ventilation; so you're 
committing a cardinal sin for non- 
invasive monitoring. You're chang- 
ing the patient-ventilator interface, 
then monitoring, and then going back 
to the original interface; so, you're 
not really monitoring the patient's 
condition for the other hours of the 
day. It works great when the patients 
are paralyzed. 

Dantzker: At least you don't have to 
worry about leaks, right? 

Branson: Oh, yeah. 

Hess: Well, leaks are a problem ei- 
ther way. 

Cerra: I want to emphasize these 
points because I think if somebody is 
getting into expired gas analysis, this 
is critical. We set up a mass spec- 
trometer system, and we came to the 
same place that you end up, too. We 
used 900Cs almost exclusively, but 
in patients with minute ventilation 




over 15 and F102S over 0.45, you can 
be off 20-25% in that closed system. 
It's a real problem, even with tech- 
nicians who do it all day long. I ad- 
vise caution in interpreting expired 
gas data. 

Wood: I want to come back to in- 
verse ratio ventilation (IRV) and the 
concept that relates mean alveolar 
pressure to oxygenation. When one 
is using IRV, it should be regarded 
as a salvage therapy. That is, once 
you"ve adjusted the patient's oxyge- 
nation on a given F102, with an 
amount of PEEP that is appropriate, 
it is possible to titrate to a little bit 
higher Po: or a little bit lower Fio 2 by 
adjusting the ventilator towards a 
higher I:E, or the IRV mode. A part 
of any beneficial effect is that the 
PEEP itself increases due to auto- 
PEEP. Conceivably, there is an addi- 
tional improvement in oxygenation 
by the mechanism whereby PEEP 
improves intrapulmonary shunt — re- 
distribution of the edema as mean 
lung volume is increased due to IRV. 
But, the incremental change over 
conventional PEEP that one gets for 
this so-called new or innovative ap- 
proach to oxygenation is 'diddley 
squat,' and only is helpful when you 
get pushed against the wall in a giv- 
en patient whom you can't manage 
with conventional levels of positive 
end-expiratory pressure. I want to 
add to that the point that often that 
approach is linked, as you did here, 
to the pressure-control mode. This is 
a 'new mode' that some people think 
diminishes barotrauma; but, just to 
get it on the record, we are sort of re- 
inventing the wheel here. This is the 
way the Bird ventilator started, and. 
to avoid barotrauma, yes, you can set 
a pressure limit or you can use a tidal 
volume of 5-7 mL/kg (which isn't as 
excessive as the 12-15 mL/kg that 
large numbers of intensivists tend to 
use) by seeking a lowest value for ti- 
dal volume that achieves adequate al- 
veolar ventilation. That's pressure- 

control ventilation. You just do it by 
turning down the tidal volume. 

The third point that I want to 
make is that it is amazing in our in- 
tensive care unit, where our respira- 
tory therapists are very well trained, 
how innovative they can become in 
the adjustments of waveforms. For 
example, the semilunar waveform, or 
the descending ramp, is just great for 
asthma, they find, because they can 
focus on reducing peak-to-plateau 
pressure because the flow at the end 
of the breath is so small with these 
waveforms that there's hardly any 
peak to pause. The problem is that 
they create large amounts of bar- 
otrauma due to auto-PEEP. So, there 
are problems with the innovations. I 
regard PCIRV as a salvage therapy 
for oxygenation. 

Branson: We rarely use PCIRV. I 
know Dr Morris is about to tell us 
differently possibly, but we do, in 
fact, have a study presented at the 
Open Forum 1 comparing volume con- 
trol with a decelerating waveform to 
pressure control in patients with 
ARDS at the same I:E, and there's 
absolutely no difference. Part of the 
problem with the PCIRV literature is 
the original studies by Tharratt et al, 2 
Lain et al, 3 and Gurevitch et al 4 were 
all poorly done with no control or re- 
porting of auto-PEEP whatsoever. I 
didn't mean to come off sounding 
like a fan. It is a way, you're right, to 
increase mean alveolar pressure, but 
I don't believe that it has any ad- 
vantage over traditional volume- 
control ventilation. 5 

1. Campbell RS, Branson RD, Johnson 
DJ, Porembka DT. Davis K Jr. Com- 
parison of pressure controlled ventila- 
tion (PCV) and volume controlled 
ventilation (VCV) with a descending 
flow waveform (abstract). Respir 
Care 1992;37(11):1317. 

2. Tharratt RS, Allen RP, Albertson TE. 
Pressure controlled inverse ratio ven- 
tilation in severe adult respiratory 
failure. Chest 1988;94:755-762. 

3. Lain DC, DiBenedetto R. Morris SL, 
Van Nguyen A, Saulters R. Causey 
D. Pressure control inverse ratio ven- 
tilation as a method to reduce peak in- 
spiratory pressure and provide ad- 
equate ventilation and oxygenation. 
Chest 1989;95:1081-1088. 

4. Gurevitch MJ, Van Dyke J. Young 
ES, Jackson K. Improved oxygena- 
tion and lower peak airway pressure 
in severe adult respiratory distress 
syndrome: treatment with inverse ra- 
tio ventilation. Chest 1986:89:211- 

5. Campbell RS, Branson RD. Ven- 
tilatory support for the '90s: pressure 
support ventilation (Kittredge's Cor- 
ner). Respir Care 1993;38(5):526- 

Morris: I certainly was pleased to 
hear Dr Wood's comments about the 
Bird ventilator. It's interesting to see 
people excited about pressure limita- 
tion. Just a few years ago people 
were excited about vol. me control 
because it wasn't pressure-limited 
ventilation. People now are excited 
about low PEEP, and a few years ago 
there was some interest in super- 
PEEP. A few years ago the ECMO 
study was criticized sound I v because 
a number of very prominent workers 
publicly declared that ARDS as a 
problem had disappeared because 
they used super-PEEP and PEEPs of 
50 cm H : had made it go away. 1 So 
the most appropriate thing some sug- 
gested was to stop the ECMO study 
before it was finished. Anyway, 
vogues become less attractive, I 
guess, as you become older. I want 
to point out one thing about the flow- 
rates of the devices that you so nice- 
ly described. All of us sitting here 
just breathing quietly have peak in- 
piratory flows of about 30 L/min; so 
it doesn't take very much for a sick 
patient to get a peak inspiratory flow 
of 60 to 80 L/min. The maximum 
I've ever measured was in burn pa- 
tients with driven ventilations related 




to the application of topical sulfon- 
amides that inhibit carbonic anhy- 
drase. They, while resting with nor- 
mal respiratory quotients (RQs) of 
about 0.85, had peak inspiratory 
flows of 180 L/min. They were clin- 
ically absolutely symptom-free, but 
looked like they had just run 10 ki- 
lometers. When manufacturers bring 
out devices that have flows of 80 or 
90 L/min, I look at them and say 
"Anybody very sick would probably 
have higher peak flows and I would 
be cautious about assuming that the 

device would control inspired frac- 
tions of oxygen." 

1. Douglas ME, Downs JB. Pulmonary 
function following severe acute res- 
piratory failure and high levels of pos- 
itive end-expiratory pressure. Chest 

Pierson: Just a quick comment about 
peak inspiratory flows. All of us, as 
we've engaged in this discussion, as 
we have inspired between sentences, 
have probably had inspiratory flows 
of well over 200 L/min, and maybe 

much more depending upon how ex- 
cited we are. 

Nelson: One of the most creative 
comments that I've heard describing 
pressure control inverse ratio ventila- 
tion was by Rocco Orlando at Hart- 
ford (personal communication) — he 
calls it surreptitious PEEP. It's for 
people who don't want to be thought 
of as using PEEP, but they want the 
same effect. And I agree, it's prob- 
ably a salvage therapy in most of our 

AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition 




American Association 
for Respiratory Care 
39th Annual Convention 
and Exhibition 

Nashville, Tennessee 

AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition • AARC 39th Annual Convention and Exhibition 



The Magic Bullets in the War on ARDS: 
Aggressive Therapy for Oxygenation Failure 

Thomas D East PhD 

I. Introduction 
II. Mechanical Ventilation 

A. Pressure Control Inverse Ratio Ventilation 

B. Airway Pressure Release Ventilation 

C. Intratracheal Gas Insufflation 

D. Independent Lung Ventilation 

E. High Frequency Ventilation 

III. Extracorporeal Gas Exchange 

A. Extracorporeal Membrane Oxygenation 

B. Extracorporeal C0 2 Removal 

C. Intravascular Oxygenator 

IV. Management Techniques 

A. Permissive Hypercapnia 

B. Expert Systems 

C. Closed-Loop Control 
V. Summary 


Adult respiratory distress syndrome (ARDS) is a 
devastating disease that can strike an otherwise vi- 
tal young adult and result in death in more than 
50% of cases. Although many new therapies have 
been introduced in the last 20 years, the reported 
survival for ARDS has remained relatively con- 

Dr East is Associate Professor of Anesthesiology, Bio- 
engineering, and Medical Informatics at the University of 
Utah, and Director of Informatics Research in the Pulmonary 
Division, LDS Hospital — Salt Lake City, Utah. 

The research on extracorporeal CO: removal was supported by 
NIH Grant HL36787 "Extracorporeal C0 2 Removal for 
ARDS," the Deseret Foundation (LDS Hospital), and the Res- 
piratory Distress Syndrome Foundation. The research on com- 
bined high frequency ventilation was supported by Siemens- 
Elema Life Support Systems. Solna, Sweden. 

The author has no financial interest in any of the products 

A version of this paper was presented by Dr East on October 8, 
1992, during the Respiratory Care Journal Conference on 
Oxygenation in the Critically 111 Patient, held in Puerto Val- 
larta. Mexico. 

Reprints: Thomas D East PhD, LDS Hospital. Pulmonary 
Division and Medical Informatics Department. 8th Ave and C 
Street, Salt Lake City UT 84143. 

stant. Growing evidence from animal and human 
studies suggests that mechanical ventilation with 
what used to be considered normal tidal volumes 
(12-15 mL/kg) can produce iatrogenic lung dam- 
age.'" 4 The mechanism for the lung damage is un- 
clear; however, evidence seems to indicate that it is 
more closely related to regional overdistention of 
the lung than to pressure. This concept of excessive 
volume led Dreyfuss et al to introduce and use the 
term "volutrauma" rather than barotrauma for this 
phenomenon. 34 As a result of this revolution in our 
understanding of the effects of mechanical ventila- 
tion, new techniques introduced into both the clin- 
ical and research environments are explicitly de- 
signed to maximize the benefit from artificial 
respiration while minimizing the harm. These new 
approaches can be grouped into those related to (1) 
mechanical ventilation, (2) extracorporeal ventila- 
tion, and (3) management techniques. 

Mechanical Ventilation 

Pressure Control Inverse Ratio Ventilation 

PCIRV is intended to reduce the likelihood of 
iatrogenic lung damage during positive pressure 
ventilation. During mechanical ventilation with 




PCIRV, peak airway pressure (Ppeak) is constant 
and the inspiratory-to-expiratory-time ratio (I:E) is 
typically 1:1 to 4:1. The inspiratory flow de- 
celerates as the airway pressure equilibrates at P pea k 
during the relatively long inspiratory time. Because 
the expiratory time (t e ) is short relative to the time 
required for complete exhalation, alveolar pressure 
may not fall to the end-expiratory pressure level ex- 
pected from the ventilator settings of PEEP. Thus, 
end-expiratory alveolar pressure (PEEPi) may be 
higher than the set PEEP. This is well recognized 5 " 8 
and known as auto-PEEP. During conventional, 
volume-limited, continuous positive pressure ven- 
tilation (CPPV) of ARDS patients in whom there is 
no auto-PEEP, adjustments of oxygenation and 
ventilation are simple and independent of one an- 
other. In contrast, for PCIRV support of ARDS. 
when auto-PEEP is present, the adjustments of oxy- 
genation and ventilation are much more complex 
and almost all adjustments of the ventilator result 
in changes in both oxygenation and ventilation. Of- 
ten these changes are in opposite directions (ie, in- 
creasing I:E may increase oxygenation and de- 
crease ventilation). These issues make PCIRV 
titration complex and confusing and underlie the 
trial and error approach commonly used with 
PCIRV in the ICU. 

Although all conclude that P pea k can be reduced 
by PCIRV and some have demonstrated decreased 
dead-space-to-tidal-volume ratio (Vd/Vt), 5 ' 9,10 in- 
creased total compliance (C to t), 5 ' 9 " increased oxy- 
gen partial pressure (PaCbX 9-12 ar >d reduced intra- 
pulmonary shunt (Qsp/Qt, 910 none of the published 
PCIRV studies is a controlled, prospective random- 
ized clinical trial focusing on outcome variables 
such as mortality or morbidity. 1314 The postulated 
mechanisms for the increased oxygenation ef- 
ficiency include recruitment of alveoli 10 " com- 
bined with a more favorable ventilation-perfusion 
ratio. 1015 Most of the authors did not find dramatic 
changes in hemodynamic variables and oxygen de- 
livery, 51617 although some reported problems. 910 
We 18 conclude from a review of the techniques 
used previously to manage PCIRV that there is no 
standard approach to the management of PCIRV, 13 
and that many articles lack the description or quan- 
tification of variables such as PEEP; and ven- 
tilatory rate (f) without which it is very difficult to 
define the conditions of the study. 

A large part of the confusion that surrounds 
PCIRV is the definition of "PEEP." This confusion 
has been propagated by a flurry of names — PEEP, 
intrinsic PEEP. auto-PEEP, inadvertent PEEP, end- 
expiratory lung pressure (EELP), volume encum- 
bered expiratory pressure (VEEP), breath stacking, 
and gas trapping — used to refer to the alveolar 
pressures generated during inverse ratio ventilation 
(IRV). 19 These terms, while sometimes used inter- 
changeably, have been associated with different 
meanings and different measurement techniques in 
different publications. While air trapping has gen- 
erally been viewed as unfavorable, some have 
pointed out the actual end-expiratory and mean 
pressure generated in the alveoli are what are im- 
portant, not how they were established. 20 " 22 They 
argue persuasively that PEEPj should be physio- 
logically equivalent to set PEEP delivered by a me- 
chanical device such as a PEEP valve or a water 
column. This assumes that changing distribution of 
PEEPi within the lung due to local time-constant 
variations is of secondary importance. 

We have developed a computerized protocol that 
provides a systematic approach for management of pre- 
ssure control inverse ratio ventilation (PCIRV). 18 
The protocols were used for 1.466 hours in 10 
around-the-clock PCIRV evaluations on 7 patients 
with severe ARDS. Patient therapy was controlled 
by protocol 95% of the time (1,396/1,466 hours), 
and 90% of the protocol instructions (1,937/2,158) 
were followed by the clinical staff. Of the 22 1 pro- 
tocol instructions, 88 (39%) not followed were due 
to invalid PEEPi measurements. Compared with 
values observed during the preceding period of 
CPPV, the expired minute ventilation was reduced 
by 27% during PCIRV, with pH not clinically dif- 
ferent (mean difference in pH = 0.02). No differ- 
ence was seen in the P a 02. PEEPi, or the F102 be- 
tween PCIRV and CPPV. The PEEP setting was 
reduced by 33% from a mean (SD) of 9 (0.5) to 6 
(0.6). and the mean (SD) inspiratory-to-expiratory 
time ratio (I:E) increased from 0.64 (0.04) to 2.3 
(0.10). Peak airway pressure was reduced by 24%, 
from 59 (1.5) to 45 (0.6), and mean (SD) airway 
pressure increased by 27%, from 22 (0.8) to 28 
(0.6) in PCIRV. Right atrial and pulmonary artery 
pressures were higher and cardiac output lower in 
PCIRV, but blood pressure was unchanged. The 
success of this protocol has demonstrated the fea- 




sibility of using PEEPi as a primary control var- 
iable for oxygenation. This computerized PCIRV 
protocol should make the future use of PCIRV less 
mystifying, simpler, and more systematic. 

Airway Pressure Release Ventilation (APRV) 

APRV is illustrated in Figure 1. A flow regulator 
and mixer provide a fresh gas flow to the patient 
through a low-resistance circuit. 23 Downstream 
from the patient there are two different threshold 
resistor valves that create two different pressure 
levels in the circuit. As the pressure is switched 
from high to low, exhalation occurs. When it is 
switched from low to high, inspiration occurs. At 
either pressure level, the patient is free to breathe 
spontaneously from the fresh gas flow. Typically, 
APRV is used with long inspiratory times and short 
expiratory times, similar to PCIRV. The primary 
distinction between APRV and PCIRV is that the 
patient can breathe spontaneously at either pressure 
level between the mechanical breaths. This has 
been proposed to aid in patient-ventilator syn- 
chrony and comfort. This mode is not commonly 
available on any commercially available ICU ven- 
tilator, and therefore little clinical experience has 
been reported in the literature. The reported draw- 
backs to APRV are similar to those of other pres- 
sure-control modes of mechanical ventilation: no 
guaranteed tidal volume, potential hemodynamic 
compromise, and risk of barotrauma. If there is no 
peak pressure limit and no pressure-relief valve, 
then a blocked expiratory pathway can result in 
rapid generation of high airway pressures from the 
continuous high flow of fresh gas. 

Fig. 1. A system for airway pressure release mechanical 
ventilation system. (Reprinted, with permission, from Ref- 
erence 23.) 

Intratracheal Gas Insufflation 

Intratracheal gas insufflation is a technique in 
which CO : elimination is improved by flushing a 
portion of the anatomic dead space with fresh gas 

(Fig. 2) 

9 x 24-27 

A blender provides the appropriate 

mixture of : and air. The flowmeter determines 
the flowrate of this gas mixture. The fresh gas flow 
is delivered through a 1.3-mm-ID catheter, placed 
with bronchoscopic guidance through the endo- 
tracheal tube 1 centimeter above the carina. This 
flow can be delivered throughout the respiratory 
cycle with flowrates above 5 L/min. If the flowrate 
is less than 5 L/min, it is more efficient to deliver 
the flush gas only during the last 33% of expira- 

Monitor Airway Pressure 
Match F102 
Reduce Vt 

2-6 L/min 

1 cm above carina 

Fig. 2. Components of an intratracheal pulmonary venti- 
lation device as reported in References 24-27. 

While the primary goal of this therapy is to im- 
prove carbon dioxide elimination, it does positively 
impact oxygenation therapy. If intratracheal gas in- 
sufflation is used, then a smaller tidal volume and 
lower peak and mean airway pressures are re- 
quired. This implies that less barotrauma might oc- 
cur to reach the same level of oxygenation. Al- 
though several animal studies have been pub- 
lished; 24 ' 21 little information is available on the ef- 
ficacy of this technique in human ARDS patients. 
The airway pressure should be monitored at the tip 
of the endotracheal tube, and it should be watched 
closely. One of the largest risks associated with this 
technique is a blocked expiratory pathway that rap- 
idly produces harmful pressure levels in the lung 
due to the continuous fresh gas flow. No systems 




for intratracheal gas insufflation equipped with 
monitors and alarms are commercially available. 

Independent Lung Ventilation (ILV) 

Independent lung ventilation has been used for 
many years for specific situations in both an- 
esthesia and critical care. 28 Independent lung ven- 
tilation is achieved by the use of endobronchial 
blockade (gauze tampon, cuffed rubber endo- 
bronchial blockers, embolectomy balloons, pul- 
monary artery catheters, urinary catheters, or spe- 
cial plastic tubes) or a double-lumen endotracheal 
tube (Robertshaw or Carlens). The most common 
indications for ILV in the ICU are bronchopleural 
fistula, asymmetrical lung disease, and airway hem- 
orrhage that cannot be managed with conventional 
mechanical ventilation. 

The biggest problem with ILV is positioning the 
tube and assuring that the tube position does not 
change. If a right main-stem bronchus double- 
lumen endobronchial tube is used, only about a 2- 
cm section of the bronchus is available in which to 
position the balloon, and the risk is high that the 
endobronchial cuff will occlude the right upper 
lobe or the carina. A left double-lumen endo- 
tracheal tube is associated with far fewer complica- 
tions due to the anatomy of the bronchial tree, 
which provides 4 to 5 cm in which to place the dis- 
tal cuff. There have been some reports that the en- 
dobronchial cuffs can be inflated to high pressures, 
damaging the tissue in the main-stem bronchus (5 
cases out of 2,700 procedures using a Carlens dou- 
ble-lumen tube). 28 

Initially many investigators felt that ventilation 
to the two lungs had to be synchronized; however, 
reports are being published now indicating little 
difference in gas exchange if the ventilators are not 
synchronized. 28 Some believe that patients may be 
more comfortable if the ventilators are synchro- 
nized. The simplest and most common approach to 
management is to use two independent ventilators 
to provide different tidal volumes and PEEP levels 
to the two lungs. In general, the tidal volume is 
smaller and the PEEP is larger in the injured lung; 
thus overdistention of the normal lung is avoided. 

High Frequency Ventilation 

High frequency ventilation has been defined as 
controlled ventilation with a tidal volume less than 

or equal to the anatomic dead space and a fre- 
quency from 60 to 3,600 breaths/min. 29 As Froese 
and Bryan point out in their review article written 
in 1987, the high frequency ventilation literature is 
confusing at best — with over 800 references in 
which almost equal numbers of authors report re- 
sults claiming positive, neutral, and negative ef- 
ficacy of high frequency ventilation. There are al- 
most as many different systems for delivering high 
frequency ventilation as there are authors. The 
most common clinically used systems are high fre- 
quency positive pressure ventilation (HFPPV), high 
frequency oscillation (HFO), high frequency jet 
ventilation (HFJV), and combined high frequency 
ventilation (CHFV). 

HFPPV uses tidal volumes of 3 to 4 mL/kg at 60 
to 100 breaths/min. Exhalation is passive. Many of 
the commercial ICU ventilators are capable of de- 
livering HFPPV at the extremes of their assist/ 
control settings. HFPPV has been primarily used 
for laryngoscopy, bronchoscopy, and upper airway 
surgery. 30 A small study in ARDS patients (n = 12) 
showed little difference in oxygenation and hemo- 
dynamics when comparing HFPPV and CPPV. 31 

HFO uses a diaphragm, bellows, or pump to ac- 
tively inspire and expire with a roughly sinusoidal 
flow waveform. Typically the respiratory fre- 
quency is from 1 to 50 Hz (60 to 3,000 breaths/ 
min). Fresh gas is added during the inspiratory part 
of the cycle, and gas is vented to the exhalation 
port during the expiratory cycle. The major ad- 
vantage of this system is the active exhalation that 
helps to avoid gas trapping associated with other 
systems that depend on passive exhalation. Al- 
though feasibility has been demonstrated for HFO. 
there are few data on the efficacy of HFO for the 
treatment of ARDS. 

The most common clinical application of HFV is 
HFJV. Gas from a high-pressure source is injected 
at periodic intervals (100-1,200 breaths/min) 
through a small cannula in the trachea. Exhalation 
is passive, depending on the elastic recoil of the 
lung and chest. There may or may not be provision 
for gas entrainment around the jet. The position of 
the jet in the trachea varies. Commercially avail- 
able endotracheal tubes have both a high and low 
port for the high frequency jet. Driving pressure, 
rate, and I:E are the primary control variables. Ti- 
dal volume is determined by the driving pressure 




and the downstream resistance and compliance. 
The biggest advantage of this system is that it is 
simple to construct and implement and provides 
high frequency breaths. The disadvantages are that 

(1) it is difficult to humidify the gas adequately and 

(2) gas trapping due to inadequate time for expira- 
tion is possible. 29 Carlon et al have shown in a 
large prospective study (n = 309) that there was no 
significant difference in outcome between ARDS 
patients supported with HFJV and those supported 
with assist-control ventilation. 32 Others have shown 
that the level of oxygenation achieved with HFJV 
is closely related to the mean airway pressure gen- 
erated. It might be that the advantage of HFJV is 
that an adequate mean airway pressure can be 
maintained without the high peak alveolar pres- 
sures associated with tidal ventilation; however, 
this hypothesis remains to be proven. 

CHFV is a mode that combines HFJV with con- 
ventional tidal ventilation. 33 The high frequency jet 
can be superimposed on the inspiratory or ex- 
piratory cycle or throughout the breath. CHFV has 
been used in a variety of human as well as animal 
trials; 33 " 43 however, the efficacy of this treatment is 
unclear due to a lack of carefully controlled ran- 
domized trials. We have conducted a randomized 
protocol-controlled animal trial of volume control 
(CPPV) and CHFV. 44 

In this trial, 20 mongrel dogs with a mean (SD) 
weight of 21 (5) kg were anesthetized, intubated, 
paralyzed, and ventilated with a Siemens 900C 
ventilator equipped with a HFV 970 unit. During 
setup, all animals were on CPPV, F102 = 0.4, Vj = 
20 mL/kg, I:E = 1 :2. Oleic acid (0.2 mL/kg) was 
infused into the pulmonary artery slowly over 15 
min. After 1 hour, the animals were randomized to 
either VC or CHF, and PEEP was set to 5 cm H 2 
in both groups. 

In the CHFV group, HFV was superimposed 
during inspiration with an initial frequency of 10 
Hz and a cycle time of 50%. The minute volume 
(Ve) was divided equally between the 900C and the 
HFV. V T of the 900C was reduced to 10 mL/kg. 

During the rest of the trial, the adjustment of 
PEEP, F102, V E (for both the 900C and HFV), rate 
of 900C, and HFV frequency was controlled by 
protocols that we have developed and tested over 5 
years in 200 ARDS patients (over 80,000 hours of 

care). The P a o2 and pH a end points were the same 
in both treatment groups. Data were collected every 
30 min for the following 6 hours. The groups were 
compared using multivariate analysis of variance 
for repeated measures. 

The protocols were successful in guaranteeing 
the same therapeutic end points in both groups, 
with mean (SD) values: Pao 2 64.5 (1.4) with CPPV, 
64.6 (1.8) with CHFV; pH a = 7.35 (0.001) with 
CPPV, 7.36 (0.00) with CHFV. The only sta- 
tistically different variables between CPPV and 
CHFV were P pea k and P paU se, which were higher in 
CHFV (F,., g , p < 0.05). 

In this animal model of ARDS, with CPPV and 
CHFV controlled carefully to provide identical pH a 
and P a o2 end points, there does not appear to be 
any additional benefit to short-term CPPV applied 
during inspiration with a frequency of 10 Hz. 

Extracorporeal Gas Exchange 

One way to 'rest' the injured lung is to remove 
part of its burden by using an extracorporeal gas 
exchanger. These extracorporeal systems consist of 
an artificial lung (typically a membrane oxy- 
genator) and a pump to help drive the blood 
through the artificial lung and back into the body. 
There are several different configurations for ex- 
tracorporeal systems. 

Extracorporeal Membrane Oxygenation 

The first reported extracorporeal systems for 
treating ARDS were introduced in 1972. 45 These 
early systems collected blood from the venous sys- 
tem and returned it to the arterial system, partially 
bypassing the natural heart and lungs. Most of the 
oxygenation and C0 2 removal was done with the 
artificial lung. This type of support is known as ex- 
tracorporeal membrane oxygenation (ECMO). 
Some membrane oxygenators have heparin -bonded 
surfaces that reduce the requirement for systemic 
heparinization and reduce the risk of bleeding. De- 
spite this advance, hemorrhage still remains the 
biggest complication of extracorporeal gas ex- 
change. Additional complications are tubing ob- 
struction, air embolization, thrombus formation, he- 
molysis, stroke, and infection. 46 A National Heart, 




Lung, and Blood Institute (NHLBI) sponsored 
prospective, randomized trial of ECMO and con- 
ventional mechanical ventilation was conducted in 
90 ARDS patients from 1974 to 1977. 47 The results 
of this trial indicated no difference in survival be- 
tween the two groups; survival = 9%. 48 ' 49 Release 
of these results halted the widespread adoption of 
ECMO as a treatment for ARDS. 

Extracorporeal CO, Removal (ECC0 2 R) 

A modification of the early ECMO systems was 
to use veno-venous bypass, whereby the blood was 
removed from the inferior vena cava and returned 
to the inferior vena cava closer to the right atrium 
(Fig. 3 50 ). 46 This eliminates the need for arterial 
cannulation and reduces the risks of bleeding, sys- 
temic thromboembolism, and hemodynamic insta- 
bility. This technique does not bypass the normal 
heart and lungs. A double-lumen percutaneous can- 
nula can be used for both the inlet and outlet to re- 
duce the complexity of cannulation. The flow 

Fig. 3. A typical ECC0 2 R system. DC = drainage cath- 
eter; RC = return catheter; ITC = intratracheal catheter 
for = 1 L/min 2 flow; ML = membrane lung; RP = roller 
pump; R = venous reservoir; H = humidifier; resp = res- 
pirator; Gl = gas inlet; GO = gas outlet; FG = gas flow 
monitor. (Reprinted, with permission, from Reference 

through the veno-venous system is lower than that 
achieved with veno-arterial bypass, and thus the 
veno-venous systems are primarily used for CO : 
removal. ECCO : R systems do provide some oxy- 
genation; however, oxygenation primarily occurs 
through the natural lung via apneic ventilation. A 
constant flow of 100% oxygen is delivered through 
a cannula into the trachea. The partial pressure gra- 
dient, typically greater than 650 torr, is enough to 
provide adequate oxygenation without any of the 
bulk gas transfer commonly associated with tidal 
ventilation. Typically, low-frequency (2-3 breaths/ 
min, Vt = 3 mL/kg) tidal ventilation is super- 
imposed on the apneic oxygenation in order to help 
prevent atelectasis. This technique is known as ex- 
tracorporeal CO : removal (ECCO : R). 51 

ECC0 2 R was first used clinically by Gattinoni 
and colleagues in the early 1980s. 5 - Their results 
from an uncontrolled clinical trial in ARDS pa- 
tients indicated a survival of 49%. This survival 
was much higher than the 9% from the ECMO 
trials of the 1970s. Several hundred patients have 
been managed with ECCO : R in Europe and the 
United States. Overall, the survival rate seems to be 
close to 50%. 46 ' 53 ' 54 

We have recently completed a NHLBI-sup- 
ported, prospective, randomized, controlled clinical 
trial designed to compare ECCO : R and conven- 
tional mechanical ventilation in patients with se- 
vere ARDS. 46,55 Computerized protocols generated 
instructions, using common logic, to yield common 
PaOz and pH a end points and to produce equal in- 
tensity of care for arterial oxygenation in all pa- 
tients. This study was conducted at the LDS Hos- 
pital in the shock trauma/respiratory ICU, a tertiary 
care ICU with an integrated electronic database and 
bedside computer terminals: 40 severe ARDS pa- 
tients who met the ECMO entry criteria were ran- 
domized into the trial. The ECMO exclusion cri- 
teria were used to ensure that enrolled patients 
would not have underlying conditions, other than 
ARDS, that would preclude survival. The main out- 
come measure was survival at hospital discharge. 
Survival was equivalent in the 19 mechanical ven- 
tilation (42%) and 21 new therapy (extracorporeal) 
patients (33%) (p = 0.8). All deaths occurred within 
30 days of randomization. Overall patient survival 
was 38% (15/40) and was about 4 times that ex- 




pected from historical data (p = 0.0002). ECC0 2 R 
survival was not significantly different from the 
published survival after extracorporeal support in 
other centers. 46 - 53 -"' 4 Mechanical ventilation survival 
was significantly higher than the 12% derived from 
published data (p = 0.0001 ). 56 Protocols controlled 
care 86% of the time; intensity of care for arterial 
oxygenation was similar in both groups (2.6 and 
2.6 PEEP changes/day; 4.3 and 5.0 F102 changes/ 

No significant difference in survival was seen 
between the mechanical ventilation and the 
ECCO : R groups. We do not currently recommend 
extracorporeal support as a therapy for ARDS. In 
our opinion, extracorporeal support for ARDS 
should be restricted to controlled clinical trials. 

Intravascular Oxygenator (I VOX) 

The IVOX is a miniaturized, elongated, hollow- 
fiber membrane oxygenator designed to be placed 
within the vena cava and to require no ex- 
travascular blood pump. 57 " 59 The IVOX (Fig. 4) is 
designed to lie within the subject's vena cava so 
that circulating venous blood can flow freely over 

Gas Outlet / vein 

Fig. 4. Intravenous oxygenator (IVOX) in place in the 
vena cava. (Reprinted, with permission, from Reference 

and around the external surfaces of the hollow fi- 
bers. Oxygen, under subatmospheric pressure, 
flows within the lumens of the hollow fibers. The 
fiber walls are covered with an ultrathin. siloxane 
membrane that permits transfer of oxygen and car- 
bon dioxide across the membrane. Water, plasma, 
solutes, and formed elements of blood do not cross 
the membrane. The IVOX is placed in the vena 
cava through a surgical venotomy in the right fe- 
moral or right jugular vein (Fig. 4). Due to the 
smaller surface area of this artificial lung, there is 
less CO : removal and oxygen uptake than would be 
seen with ECMO or ECC0 2 R. 60 IVOX is intended 
for temporary (up to 21 days) augmentation of gas 
exchange in ARDS patients with moderate blood 
gas deficits. The advantage of IVOX is that there is 
no need for an extracorporeal circuit and pump, 
which is considerably simpler and safer than the 
use of ECMO or ECCO.R. 61 

Clinical trials of IVOX are being conducted un- 
der FDA supervision. Phase- 1 trials demonstrating 
safety of IVOX have been completed. Phase-2 
trials, assessing the efficacy of the device, are cur- 
rently under way at 18 clinical centers throughout 
the world. 59 - 6264 As of August 1, 1991, 56 patients 
had been cared for using the IVOX device; how- 
ever, data were complete in only 36 patients. 59 
Based on their Murray scores, 65 these patients had 
an expected survival of 20%. Eleven of the 36 
IVOX patients survived (30%). Significant benefit 
(defined as "blood gases have been improved, me- 
chanical ventilator parameters have been reduced, 
improved lung injury scores, improved hemo- 
dynamics) has been observed in 85% of the IVOX 
patients. No significant complications, problems, or 
unfavorable sequelae attributable to IVOX utiliza- 
tion have been recognized clinically or by clinical 
laboratory tests in any of the first 36 patients. No 
significant gross or histologic pathologic findings 
attributable to IVOX utilization have been ob- 
served at the first 14 necropsy examinations of 
ARDS patients dying after IVOX utilization. The 
results on the IVOX device to date indicate that it 
appears to be safe; however, there is little con- 
clusive evidence as to its efficacy in treating ARDS 
as compared to conventional mechanical ventila- 
tion. At this time, IVOX should only be used in the 
context of a clinical trial. 




Management Techniques 

Permissive Hypercapnia 

Although some may view permissive hyper- 
capnia as a mode of ventilation all its own, it is 
really more of a management style that can be ap- 
plied to virtually all the modes, mechanical and 
extracorporeal, of ventilation. The concept is to 
minimize the risk of barotrauma and volutrauma by 
using the smallest volumes and the lowest pres- 
sures that provide adequate gas exchange. 6667 In 
order to do this in ARDS patients, one can com- 
promise on C0 2 elimination. The goal is to keep 
the arterial pH low (between 7.25 and 7.35), use a 
small tidal volume (6-8 mL/kg), and avoid high 
peak and pause airway pressures (keep pause pres- 
sure below 45 cm H 2 0). There is strong evidence 
from animal work 1,3,4 and suggestive evidence from 
a prospective human trial 68 that using permissive 
hypercapnia as a management style may help to 
avoid barotrauma or volutrauma. No prospective 
randomized trials have been conducted to de- 
termine the efficacy of permissive hypercapnia. 

Expert Systems 

Expert systems are decision support tools that 
are designed to aid the clinician at the bedside with 
ventilator management. A number of different ex- 
pert systems have been developed for ventilator 
management. 69 " 74 but few have been accepted into 
routine use in the clinical environment. 

Stimulated by investigative needs of a recent 
ECC0 2 R clinical trial, 55 we developed protocols to 
standardize therapy. 75 ' 76 We reasoned that standard- 
ization of therapy would increase the inter- 
pretability and credibility of our clinical trial re- 
sults. 55 Our protocol-control goals were to ensure 
uniformity of care, equal intensity and frequency of 
monitoring, consistent decision-making logic, and 
common therapeutic goals (eg, PaO:)- We de- 
veloped protocols for CPPV, PCIRV, low- 
frequency positive pressure ventilation with ex- 
tracorporeal CO : removal (LFPPV-ECC0 2 R), 
IMV, and CPAP for application in patients with 
ARDS. For PCIRV only, a protocol for ventilation 
and pH a control was also used. 18 

From the time patients were randomized until 
their extubation or death, we attempted around-the- 
clock bedside protocol application, for 19,920 total 
hours in the 40 patients. Of these 19,920 total 
hours, 36% was with paper-based flow diagrams in 
the first 1 8 randomized patients, and 64% was with 
the computerized HELP system protocols in the 
last 22 randomized patients. 

The protocols controlled management of arterial 
oxygenation for 17,206 hours (86% of the 19,920 
hours). Arterial oxygenation management was con- 
trolled in the control patient group for 8,402 of 
9,446 hours (89%) and in the new therapy group 
for 8,804 of 10,473 hours (84%). In the 22 patients 
supported with computerized protocols, 10,300 of 
11,712 (88%) protocol therapy instructions gener- 
ated were followed by the clinical staff. 

In summary, protocol control of severely ill ICU 
patients seems feasible. A satisfactory computer in- 
frastructure makes protocol control practical. Our 
observed four-fold survival increase associated 
with protocol control suggests that it is not harm- 
ful. Protocol control represents a medical decision 
support approach for standardizing therapy. By 
standardizing therapy, protocols are likely to sig- 
nificantly reduce clinical environment noise and re- 
duce bias, especially important in clinical trials that 
cannot be blinded. Interpretation of outcomes re- 
search results should thereby be made easier and 
conclusions should be more credible and more like- 
ly to contribute to medical policy formulation. 

Closed-Loop Control 

Many different systems has been designed over 
the years for closed-loop control of mechanical 
ventilation; 7778 however, none of these has had a 
major impact on clinical care, I believe primarily 
for two reasons. (1) These systems rely on input 
data from sensors (eg, P a c>2> S a 02) that are too un- 
reliable to trust in closed-loop control. (2) Most of 
these systems were designed by engineers in the la- 
boratory; and while they are excellent engineering 
exercises, they are not closely related to common 
clinical practice. 

We have recently tried to address this problem 
by designing a system based on well-established 
protocols for management of mechanical ventila- 






PEEP, FI02, 

Both or Neither 

Should Be 


Using 5th 


Quality Surface 


Current PEEP, Current FI02 


K0(e P0+e1 P1+e2P2) 




K0(eF0+e1F1+e2 F 2) 

Fig. 5. A closed-loop controller for ad- 
justing PEEP and F102. The general con- 
cept is that the PEEP and F102 are auto- 
matically adjusted to maintain P a o2 at a 
desired setpoint. The size and direction 
of a therapy change were determined by 
the P a 02 error, "e" (difference between 
measured and desired P a 02>- A record is 
kept of the values previous two for the er- 
ror (e1 and e2). The decision as to which 
variable, PEEP or F102 or both, to change 
was determined from a fifth dimensional 
quality surface that described the trade- 
off between PEEP and F102 changes for 
a specific combination of current PEEP 
F102, and P a o2 values. The quality sur- 
face was generated from the clinically 
proven protocols. Once it has been de- 
cided which variables are to be changed 
the amount of change is calculated by a 
proportional, integral, and differential 

(PID) controller. The controller uses the current (e) and previous values (e1 and e2) to calculate the actual change in PEEP 
or Fi02- The PID gain constants PO, P1, P2, FO, F1, and F2 are tuned to provide the best controller performance. The over- 
all gain of the controller (KO) was a function of the S a o 2 . This provided an aggressive response to hypoxemia and a more 
conservative response to hyperoxia. Arterial oxygen saturation (S a 02) was calculated from the P a o2 (assuming pH = 7.4, 
PaC02 = 35, and temperature = 37°C) using standard equations. 


Sa02=f(Pa02) Assume Standard: 
T=37°C,PaC02=35, and pH=7.4 

K0=[2' 60 ( 1 02-SaO2)]/900,000 

tion, 75 ' 76,79 which provides continuous closed-loop 
control of oxygenation (Fig. 5). 80 A Hamilton 
Amadeus ventilator was controlled by an Apple 
Macintosh SE-30 computer (Apple Computer Inc, 
Cupertino CA). A Shiley Continucath intra-arterial 
electrode (Shiley Inc, Irvine CA) was used to pro- 
vide continuous measurement of P a o:- Arterial oxy- 
gen saturation (S a o:) was calculated from the P a c>2< 
pH, PaCO:. and temperature using standard equa- 
tions. The size and direction of a therapy change 
were determined by the P a 02 error (difference be- 
tween measured and desired P a o:)- The overall gain 
of the controller was a function of the SaOz- This 
provided an aggressive response to hypoxemia and 
a more conservative response to hyperoxia. The 
content of the therapy change was determined from 
a fifth dimensional quality surface that described 
the tradeoff between PEEP and F102 changes for a 
specific combination of current PEEP, Fich. and 
P a 02 values. The quality surface was generated 
from the clinically proven protocols. The controller 
was tested in five mongrel dogs who had an oleic 
acid lung injury. After injury, the controller was ac- 
tivated and the animals were monitored at 30- 

minute intervals for 6 hours. Figure 6 illustrates the 
controller performance for one particular animal. 
The controller performed as designed for all five 
animals, and no hazardous or failure conditions 
were noted. We hope that this careful management 
of potentially toxic therapies will improve outcome 
in patients with hypoxic respiratory failure. 

1811 240 300 360 420 4X0 

Fig. 6. Typical response of the closed-loop controller in 
an animal model of ARDS. 




Table 1 . Summary of New Techniques and the Clinical Studies Done To Determine Efficacy 


Clinical Studies 


Mechanical Ventilation 



Extracorporeal Ventilation 


Management Techniques 

Permissive hypercapnia 

Expert systems 
Closed-loop control 

Many small retrospective and prospective trials, few controlled Perhaps 

Some prospective trials, not randomized or controlled Perhaps 

Case reports ? 

Case reports and in specific clinical situations Yes. for 

specific cases 
Many small prospective studies, few randomized and controlled No to Perhaps 

90-patient prospective, randomized, controlled trial No 

Many small studies (prospective, uncontrolled, nonrandomized) and one 40-patient. No 

prospective, randomized, controlled clinical trial 

Ongoing prospective, uncontrolled, nonrandomized trial Perhaps 

Case reports, small prospective, uncontrolled trials, and one prospec- Perhaps 

tive, randomized, controlled trial in planning stage 
Prospective, uncontrolled trial and one ongoing, prospective, randomized. Perhaps 

controlled trial 
Small prospective, uncontrolled trials ? 


Table 1 summarizes the new "silver bullets" in 
our armamentarium for the battle against ARDS. I 
believe that the efficacy of most of these tech- 
niques is questionable at best. The one thing that 
seems to be settling out from these different tech- 
niques is that mechanical ventilation is harmful and 
should be titrated as carefully as possible to max- 
imize the benefit and minimize the harm. It may be 
that the management process of any of the mo- 
dalities is far more important than the choice of 
ventilation technique. It is clear that we need more 
carefully controlled, prospective randomized trials 
of these new techniques during which the two 
groups of subjects receive standardized care to as- 
sure equal intensity of care and common ther- 
apeutic end points. Although such studies are la- 
borious and expensive, I believe that this is the 
only way that we will ever learn which, if any, of 
these silver bullets are really capable of making an 
impact on ARDS. 


1. Kolobow T, Moretti MP, Fumagalli R. Mascheroni D. 
Prato P. Chen V, et al. Severe impairment in lung func- 
tion induced by high peak airway pressure during me- 

chanical ventilation: an experimental study. Am Rev 
RespirDis 1987;135:312-315. 

2. Tsuno K, Prato P, Kolobow T. Acute lung injury from 
mechanical ventilation at moderately high airway pres- 
sures. J Appl Physiol 1 990:69(3 ):956-961. 

3. Dreyfuss D, Basset G. Soler P, Saumon G. Intermittent 
positive-pressure hyperventilation with high inflation 
pressures produces pulmonary microvascular injury in 
rats. Am Rev Respir Dis 1985:132:880-884. 

4. Dreyfuss D, Soler P, Basset G. Saumon G. High in- 
flation pressure pulmonary edema: respective effects of 
high airway pressure, high tidal volume, and positive 
end-expiratory pressure. Am Rev Respir Dis 1988:137: 

5. Baum M, Benzer H, Mutz N, Pauser G, Tonczar L. In- 
versed ratio ventilation (IRV): die rolle des atemzeitver- 
haltvisses in der veatmung beim ARDS. Anaesthesist 

6. Rossi R, Gottfried SB, Zocchi L. Higgs BD. Lennox S. 
Calverley PM, et al. Measurement of static compliance 
of the total respiratory system in patients with acute res- 
piratory failure during mechanical ventilation: the effect 
of intrinsic positive end-expiratory pressure. Am Rev 
RespirDis 1985;131:672-677. 

7. Caviedes I. Benito S, Mancebo J. Net A. Effect of in- 
trinsic positive end-expiratory pressure on respiratory 
compliance. Crit Care Med 1986;14:947-949. 

8. Pepe PE. Marini JJ. Occult positive end-expiratory pres- 
sure in mechanically ventilated patients with airflow ob- 
struction: the auto-PEEP effect. Am Rev Respir Dis 




9. Cole AG, Weller SF. Sykes MK. Inverse ratio ventila- 
tion compared with PEEP in adult respiratory failure. 
Intensive Care Med 1984;10:227-232. 

10. Tharratt RS, Allen RP, Albertson TE. Pressure con- 
trolled inverse ratio ventilation in severe adult res- 
piratory failure. Chest 1988;94:755-762. 

1 1 . Ravizza AG, Carugo D, Cerchiari EL, Ferrante M, Delia 
Pupa T, Villa L. Inversed ratio and conventional ventila- 
tions: comparison of the respiratory effects (abstract). 
Anesthesiology 1983;59:A523. 

12. Lain DC, DiBenedetto R, Morris SL, Van Nguyen A, 
Saulters R, Causey D. Pressure control inverse ratio 
ventilation as a method to reduce peak inspiratory pres- 
sure and provide adequate ventilation and oxygenation. 
Chest 1989;95:1081-1088. 

13. Kacmarek RM, Hess D. Pressure-controlled inverse- 
ratio ventilation: panacea or auto-PEEP? (editorial). Re- 
spirCare 1990;35:945-948. 

14. Perel A. Newer ventilation modes — temptations and pit- 
falls (editorial). Crit Care Med 1987;15:707-709. 

15. Gattinoni L. Pesenti A, Caspani ML, Pelizzoli A. Mas- 
cheroni D, Marcolin R, et al. The role of total static lung 
compliance in the management of severe ARDS un- 
responsive to conventional treatment. Intensive Care 
Med 1984;10:121-126. 

16. Lachmann B, Danzmann E, Haendly B, Jonson B. Ven- 
tilator settings and gas exchange in respiratory distress 
syndrome. In: Prakash O, ed. Applied physiology in 
clinical respiratory care. Boston: Martinus Nijhoff, 

17. Abraham E, Yoshihara G. Cardiorespiratory effects of 
pressure controlled inverse ratio ventilation in severe 
respiratory failure. Chest 1989;96:1356-1359. 

18. East TD, Bohm SH, Wallace CJ, Clemmer TP, Weaver 
LK, Orme JF Jr, et al. A successful computerized proto- 
col for clinical management of pressure control inverse 
ratio ventilation in ARDS patients. Chest 1992:101:697- 

19. Lain DC, Chaudhary BA, Thorarinsson B, Speir WA Jr. 
Auto-PEEP and proximal airway pressure: need for clar- 
ification (editorial). Chest 1990:97:771. 

20. Benson MS, Pierson DJ. Auto-PEEP during mechanical 
ventilation of adults. RespirCare 1988;33:557-568. 

21. Duncan SR, Rizk NW, Raffin TA. Inverse ratio ventila- 
tion: PEEP in disguise? (editorial). Chest 1987;92:390- 

22. Walley KR. Schmidt GA. Therapeutic use of intrinsic 
positive end-expiratory pressure. Crit Care Med 1990; 

23. Stock MC, Downs JB. Airway pressure release ventila- 
tion: a new approach to ventilatory support during acute 
lung injury. RespirCare 1987;32:517-524. 

24. Aprigliano M, Kolobow T, Rossi N. Tsuno K, Lee KH. 
Intratracheal pulmonary ventilation (ITPV) in the man- 
agement of acute respiratory failure (ARF) (abstract). 
Am Rev Respir Dis 1992;145(4, Part 2):A455. 

25. Burke WC, Nahum A, Ravenscraft SA, Adams AB, Na- 
kos G, Marcy TW, et al. Phase specific intratracheal gas 
insufflation augments alveolar ventilation (abstract). 
Am Rev Respir Dis 1992;145(4, Part 2):A455. 

26. Nahum A, Burke WC, Ravenscraft SA, Marcy TW. Ad- 
ams AB, Nakos G, et al. Augmentation of CO: elimina- 
tion by intratracheal gas insufflation during pressure 
control ventilation in dogs (abstract). Am Rev Respir 
Dis 1992;145(4, Part 2):A529. 

27. Ravenscraft SA, Burke WC, Nahum A, Marcy TW, Ad- 
ams AB, Nakos G, et al. Intratracheal gas insufflation 
augments alveolar ventilation during volume cycles me- 
chanical ventilation in patients (abstract). Am Rev Re- 
spir Dis 1992; 145(4, Part 2):A529. 

28. Brodsky JB. Mihm FG. Split-lung ventilation. In: Hall 
JB, Schmidt GA, Wood LD, eds. Principles of critical 
care. New York: McGraw-Hill, 1992:160-164. 

29. Froese AB, Bryan AC. High frequency ventilation. Am 
Rev Respir Dis 1987;135:1363-1374. 

30. Sjostrand U. High-frequency positive-pressure ventila- 
tion (HFPPV): a review. Crit Care Med 1980,8:345- 

31. Wattwil LM, Sjostrand UH, Borg UR. Comparative 
studies of IPPV and HFPPV with PEEP in critical care 
patients. I: a clinical evaluation. Crit Care Med 1 983; 1 1: 

32. Carlon GC, Howland WS, Ray C, Miodownik S, Griffin 
JP, Groeger JS. High-frequency jet ventilation: a pros- 
pective randomized evaluation. Chest 1983;84:551-559. 

33. El-Baz N, Faber LP, Doolas A. Combined high- 
frequency ventilation for management of terminal respi- 
ratory failure: a new technique. Anesth Analg 1983:62: 

34. Lachmann B. Schairer W. Hafner M, Armbruster S, Jon- 
son B. Volume-controlled ventilation with superim- 
posed high frequency ventilation during expiration in 
healthy and surfactant-depleted pig lungs. Acta Anaes- 
thesiol Scand Suppl 1989;33(Suppl 90): 1 17-1 19. 

35. Blum-Hoffmann E. Kopotic RJ, Mannino FL. High- 
frequency oscillatory ventilation combined with inter- 
mittent mandatory ventilation in critically ill neonates: 3 
years of experience. Eur J Pediatr 1988;147:392-398. 

36. Brichant JF, Rouby JJ, Viars P. Intermittent positive 
pressure ventilation with either positive end-expiratory 
pressure or high-frequency jet ventilation (HFJV). or 
HFJV alone in human acute respiratory failure. Anesth 
Analg 1986;65:1135-1142. 

37. Obara H, Maekawa N, Hamatani S, Iwai S. Effect of 
combined high-frequency oscillatory ventilation on rab- 
bit alveolar type-2 cells. Crit Care Med 1987;15(7):692- 

38. Spitzer AR, Butler S, Fox WW. Ventilatory response to 
combined high frequency jet ventilation and conven- 
tional mechanical ventilation for the rescue treatment of 
severe neonatal lung disease. Pediatr Pulmonol 1989;7: 




39. Zeravik J, Eckart J, Zimmermann G, Blumel G, Pfeiffer 
UJ, Wellhofer H. Indications for combined high fre- 
quency ventilation in clinical use. Acta Anaesthesiol 
Scand Suppl 1989;33(Suppl 90):149-152. 

40. Borg UR. Belzberg H. Blevins S. Combined high fre- 
quency ventilation (CHFV). Acta Anaesthesiol Scand 
Suppl 1989;33(Suppl90):155-157. 

41. Borg UR, Stoklosa JC. Siegel JH. Wiles CE 3rd, Belz- 
berg H, Blevins S, et al. Prospective evaluation of com- 
bined high-frequency ventilation in post-traumatic pa- 
tients with adult respiratory distress syndrome refractory 
to optimized conventional ventilatory management. Crit 
Care Med 1989;17:1129-1142. 

42. Andersen JB. Ventilatory strategy in catastrophic lung 
disease. Inversed ratio ventilation (IRV) and combined 
high frequency ventilation (CHFV). Acta Anaesthesiol 
Scand Suppl 1989;33(Suppl 90): 145-148. 

43. Zeravik J, Pfeiffer UJ. Efficacy of high frequency ven- 
tilation combined with volume controlled ventilation in 
dependency of extravascular lung water. Acta Anaes- 
thesiol Scand Suppl 1989;33(Suppl 90):568-574. 

44. East TD. Danino D, Pace NL. A randomized protocol- 
controlled trial of volume control and combined high 
frequency ventilation in a canine model of ARDS. Crit 
Care Med 1992:20(4):S62. 

45. Hill JD. O'Brien TG, Murray JJ, Dontigny L, Bramsom 
ML, Osbom JJ, et al. Prolonged extracorporeal oxy- 
genation for acute post-traumatic respiratory failure 
(shock-lung syndrome): use of the Bramsom membrane 
lung. N Engl J Med 1972;286:629-634. 

46. Sznajder JI, Morris AH. Extracorporeal membrane oxy- 
genation and CO2 removal. In: Hall JB. Schmidt GA, 
Wood LDH. eds. Principles of critical care. New York: 
McGraw-Hill, 1992:374-382. 

47. NIH-HLBI. Protocol for extracorporeal support for res- 
piratory insufficiency collaborative program. Bethesda: 
National Heart & Lung Institute. Division of Lung Dis- 
eases, 1974. 

48. Zapol WM, Snider MT. Hill JD. Fallat RJ. Bartlett RH, 
Edmunds LH, et al. Extracorporeal membrane oxy- 
genation in severe acute respiratory failure: a ran- 
domized prospective study. JAMA 1979:242:2193- 

49. NIH-HLBI. Extracorporeal support for respiratory in- 
sufficiency a collaborative study in response to RFP- 
NHLI-73-20. Bethesda: U.S. Department of Health, Ed- 
ucation and Welfare. National Institutes of Health. 

50. Morris AH, Wallace CJ, Clemmer TP, Orme JF Jr. 
Weaver LK. Dean NC, et al. Extracorporeal CO: re- 
moval therapy for adult respiratory distress syndrome 
patients: computerized protocol controlled trial. Rean 
Soins intens Med Urg 1990:6(7)485-490. 

51. Morris AH, Wallace CJ, Clemmer TP, Orme JF Jr. 
Weaver LK, Dean NC, et al. Extracorporeal CO: re- 
moval therapy for adult respiratory distress syndrome 
patients. RespirCare 1990:35:224-231. 

52. Gattinoni L, Pesenti A, Mascheroni D. Marcolin R, Fu- 
magalli R, Rossi F. et al. Low-frequency positive- 
pressure ventilation with extracorporeal CO: removal in 
severe acute respiratory failure. JAMA 1986:256:881- 

53. Pesenti A. Gattinoni L, Kolobow T, Damia G. Extra- 
corporeal circulation in adult respiratory failure. ASAIO 
Trans 1988:34:43-47. 

54. Wagner PK. Knoch M, Sangmeister C, Muller E, Len- 
nartz H, Rothmund M. Extracorporeal gas exchange in 
adult respiratory distress syndrome: associated morbid- 
ity and its surgical treatment. Br J Surg 1990:77:1395- 

55. Morris AH, Wallace CJ. Clemmer TP, Orme JF, Weav- 
er LK, Thomas F, et al. Final report: computerized pro- 
tocol controlled clinical trial of new therapy which in- 
cludes ECC0 2 R for ARDS (abstract). Am Rev Respir 
Dis 1992:145(4, Part 2):AI84. 

56. Zapol WM, Frikker MJ. Pontoppidan H, Wilson RS, 
Lynch KE. The adult respiratory distress syndrome at 
Massachusetts General Hospital — etiology progression 
and survival rates, 1978- 1988. In: F Walz, ed. Adult res- 
piratory distress syndrome. New York: Marcel Dekker 
Inc, 1991:367-380. 

57. Mortensen JD, Berry G. Conceptual and design features 
of a practical, clinically effective intravenous mechan- 
ical blood oxygen/carbon dioxide exchange device 
(IVOX). Int J Artif Organs 1989;12:384-389. 

58. Mortensen JD. Augmentation of blood gas transfer by 
means of an intravascular blood gas exchanger (IVOX). 
In: Marini JJ. Roussos C, eds. 1991 Update in intensive 
care and emergency medicine: ventilatory failure. Vol 
15. New York: Springer-Verlag, 1991:318-346. 

59. Mortensen JD. Intravascular oxygenator: a new alter- 
native method for augmenting blood gas transfer in pa- 
tients with acute respiratory failure. Artif Organs 1992; 

60. Bagley B. Bagley A. Henrie J, Froerer C. Brohamer J, 
Burkart J. et al. Quantitative gas transfer into and out of 
circulating venous blood by means of an intravenacaval 
oxygenator. ASAIO Transactions 1991;37:M413-M415. 

61. Durbin CJ Jr. Intravenous oxygenation and CO: re- 
moval device: IVOX. RespirCare 1992;37:147-153. 

62. von Segesser LK, Schaffner A, Stocker R, Lachat M, 
Speich R, Baumann PC, et al. Extended (29 days) use of 
intravascular gas exchanger (letter). Lancet 1992:339 
(June 20): 1536. 

63. Lanigan CJ. Withington PS. Support when gas ex- 
change fails— ECMO, ECCO2R and IVOX. Clin Int 
Care 1991;2(4):210-216. 

64. Kallis P, al-Saady NM. Bennett ED, Treasure T. Intra- 
vascular oxygenation with the IVOX. Br J Hosp Med 
1992:47(1 1):824-828. 

65. Murray JF, Matthay MA, Luce JM, Flick MR. An ex- 
panded definition of the adult respiratory distress syn- 
drome. Am Rev Respir Dis 1988;120:720-725. 




66. Hickling KG, Henderson SJ, Jackson R. Low mortality 
associated with low volume pressure limited ventilation 
with permissive hypercapniaTrrsevere adult respiratory 
distress syndrome. Intensive Care Med 1990; 16:372- 

67. Hickling KG. Ventilatory management of ARDS: can it 
affect the outcome? Intensive Care Med 1990;16:219- 

68. Kacmarek RM, Hickling KE. Permissive hypercapnia. 
RespirCare 1993;38:373-387. 

69. Miller PL. Goal oriented critiquing by computer for 
ventilatory management. Comput Biomed Res 1985:18: 

70. Menn SJ, Barnett GO, Schnechel D, Owens WD, Pon- 
toppidan H. A computer program to assist in the care of 
acute respiratory failure. JAMA 1973;223:308-312. 

71. Grossman R, Hew E, Aberman A. Assessment of the 
ability to manage patients on mechanical ventilators us- 
ing a computer model. Acute Care 1984;10:95-102. 

72. Fagan L, Kunz J, Feigenbaum E, Osborn J. Repre- 
senting time -dependent relations in a medical setting 
(PhD dissertation). Stanford CA: Stanford University, 

73. East TD. Henderson S, Morris AH, Gardner RM. Imple- 
mentation issue and challenges for computerized clin- 
ical protocols for management of mechanical ventilation 
in ARDS patients. In: Proceedings of the symposium on 
computer applications in medical care (SCAMC), Nov 
5-8, 1989. Washington DC: IEEE Computer Society 
Press, 1989:583-587. 

74. Henderson S, East TD, Morris AH, Gardner RM. Per- 
formance evaluation of computerized clinical protocols 

for management of mechanical ventilation in ARDS pa- 
tients. In: Proceedings of the symposium on computer 
applications in medical care (SCAMC). Nov 5-8, 1989. 
Washington DC: IEEE Computer Society Press, 1989: 

75. East TD, Henderson S, Pace NL, Morris AH. Brunner 
JX. Knowledge engineering using retrospective review 
of data: a useful technique or merely data dredging? Int 
J Clin Monit Comput 1991-1992:8:259-262. 

76. East TD, Morris AH, Wallace CJ, Clemmer TP, Orme 
JF Jr, Weaver LK. et al. A strategy for development of 
computerized critical care decision support systems. Int 
J Clin Monit Comput 1991-1992:8:263-269. 

77. Tehrani FT. A microcomputer oxygen control system 
for ventilatory therapy. Ann Biomed Eng 1992:20:547- 

78. East TD, in't Veen JC, Jonker TA, Pace NL, McJames 
S. Computer-controlled positive end-expiratory pressure 
titration for effective oxygenation without frequent 
blood gases. Crit Care Med 1988:16(3):252-257. 

79. Henderson S. Crapo RO, Wallace CJ, East TD, Morris 
AH, Gardner RM. Performance of computerized proto- 
cols for the management of arterial oxygenation in an 
intensive care unit. Int J Clin Monit Comput 1991-1992; 

80. East TD, Tolle CR, Farrell RM, Brunner JX. A non- 
linear closed-loop controller for oxygenation based on a 
clinically proven fifth dimensional quality surface. Crit 
Care Med 1991;19(4):S61. 

East Discussion 

Hudson: Were you or Alan (Morris) 
involved in the IVOX trial, 1 and 
what was the source of your data for 
the IVOX? 

1. Mortensen JD. Conceptual and design 
features of a practical, clinically ef- 
fective intravenous mechanical blood 
oxygen/carbon dioxide exchange de- 
vice (IVOX). Int J Artif Organs 1989; 

East: The IVOX data were given to 
me by JD Mortensen, one of the in- 
ventors of IVOX, and so may be 
biased. 1 

1. Bagley B, Bagley A, Henrie J, Froer- 
er C, Brohamer J. Burkart J, Morten- 
sen JD. Quantitative gas transfer into 

and out of circulating venous blood 
by means of an intravenacaval oxy- 
genator. ASAIO Trans 1991:37(3): 


Hudson: The point here is that I 
think the data you showed could be 
misleading. One of our services has 
been participating in the trial. The 
chief started out as a real advocate of 
this, but we have had significant 
complications and potentially ad- 
verse physiologic changes. They 
don't necessarily get reported be- 
cause they aren't asked about as part 
of the protocol. I was appalled when 
I heard a presentation of the pre- 
liminary data. For example, they 
aren't looking at what happens to the 
cardiac output when this device is 
put in. We have seen major falls in 
cardiac output and yet it isn't re- 

ported; so, there are 'no complica- 
tions.' I really think it brings up one 
of the problems. Not only do they 
need randomized controlled trials, 
but also we as investigators have to 
be involved in developing the proto- 
col, and we have to make sure that 
the relevant observations are made 
and reported. 

East: I agree with you. I asked Dr 
Mortensen for a collection of papers, 
and (in all fairness to him) he gave 
me a variety of papers, even ones he 
hadn't written. But, of course, they 
all tend to emphasize the positive. 
My conclusions were similar to 
yours that, if anything, the data they 
have on efficacy are very soft. 

Hudson: One of the problems, so 
far, is that most of the centers, most 




of the countries in fact, have 1 or 2 
patients, and of the 12 patients from 
the United States, I think most of 
them are from 3 centers. So, most 
people have experience with 1 pa- 
tient. I really don't think we should 
be enthusiastic about this yet. 

Morris: We've done 1 patient, I 
think. Did Terry Clemmer put in the 

East: Yes, he did the first one during 
our ECCO2R trial around 1987-1988. 

Morris: So, we inserted the first one. 
That's the only one we've ever done. 
We've encouraged the company to 
conduct randomized, controlled 
trials, and offered to participate and 
develop protocols that could be used 
on multiple centers. It's a clever de- 
vice. It will be a shame if it isn't ef- 
fectively evaluated. I hope it will 
happen. Perhaps, if you've been one 
of the centers, we could reinforce 
that need. 

Wood: A couple of interrelated com- 
ments on the data that you presented: 
The controlled hypoventilation trial 
is going to reveal some surprising re- 
sults. The 6-8 mL/kg tidal volume is 
going to make a lot of intensivists 
wonder why they can't make their 
patients hypoventilate. We use 6-8 
mL/kg as our standard tidal volume 
in ARDS, and maintain normal CO? 
level. When we alter a time-honored 
tradition such as tidal volumes of 12- 
15 mL/kg to make an 'aggressive' in- 
novation, we should first ac- 
knowledge that the effective part of 
the novelty is in rejecting excessive 
intervention to achieve the thera- 
peutic goal. Using a smaller tidal 
volume appropriate to lung size is 
hardly 'aggressive.' Similarly, high 
frequency jet ventilation, intra- 
tracheal oxygen instillation, and high 
frequency oscillation are systems in 
which you can't measure mean air- 
way pressure unless you do some 
complicated tricks to stop the ven- 
tilator and to seal the airway si- 

multaneously; without this interven- 
tion, the Bernoulli effect causes you 
to underestimate by 5-7 cm H2O 
what that mean airway pressure is. 
Because the rationale for the use of 
those salvage devices in oxygen fail- 
ure is to re-expand alveoli and re- 
distribute the edema, and thereby de- 
crease the shunt, it isn't a surprise 
that the oxygenation improves. It fre- 
quently does in those many papers 
that you told us about. But the reason 
for any benefit is the PEEP; the word 
'surreptitious' PEEP comes back. 
Again, people don't want to use 
PEEP, so they use a high frequency 
oscillator and now they can't see the 
PEEP because they can't measure 
the mean airway pressure. The skep- 
tical approach to each of these de- 
vices is just so welcome in a group 
of people like this (I hope skepticism 
is as prevalent in the wide com- 
munity of critical care) because what 
thoughtful intensivists should do is 
to employ the innovative device or 
approach as an extension of the con- 
ventional rationale or the therapeutic 
objectives rather than to try a mag- 
ical device because nothing else 
works. I advertise our critical care 
group as having high frequency os- 
cillators, and so we get called to see 
the difficult patient. Most often in- 
stead of providing the oscillator, I 
usually end up recommending a re- 
duction in tidal volume, an increase 
in PEEP, and small reductions in the 
Fio;, and thus minimize the adverse 
cardiopulmonary interactions that are 
there. That is, implementing goals of 
therapy is a lot more complicated 
than hooking up a noisy, dangerous 
ventilator, but I believe it's a lot 
more efficacious. 

East: Yes. 

Wood: That's not what everybody 
believes. Many think that critical 
care is a compendium of tech- 
niques — just choose and implement 
the best and your patient has a better 
chance to recover. I think that ap- 

proach to critical care technology 
should be superseded by the in- 
tensivists' intellectual engagement in 
a more productive process of prob- 
lem-finding and problem-solving. 
Once so engaged, the physician can 
respond creatively to the needs of the 
whole patient rather than be confined 
by a series of directives prepared by 
experts who are not at the bedside. 
I'll bet that all of us, if we set out to 
do so, could provide considerable in- 
put on Alan's (Morris) choices for 
his expert directives that are going to 
go out. The algorithms might not let 
us get out of this room tonight while 
we adjusted them. On the other hand, 
even that effort is more likely to mis- 
lead than to guide intensivists who 
choose to follow the algorithms — 
unless they are intellectually engaged 
in defining the therapeutic goals for 
their patients as opposed to imple- 
menting thoughtful directives. 

East: I shouldn't give you the im- 
pression that Alan (Morris) did it 
alone — it was a group of physicians 
from not only our institution, but 
also from Gattinoni's group and oth- 
er people. It was a consensus group 
that did the logic. 12 But you're still 
right. Even if it's 10 or 15 phy- 
sicians, if I presented it to any one of 
you or even one of the 10 or 15, it's 
different from how they would do it. 
Dr Morris has used the example of 
baking a cake. If you use a recipe, 
it'll reliably tum out a good cake 
every time, but that doesn't mean the 
master chef can't produce a much 
finer cake once in a while. But, on 
the whole, if you use a recipe, you'll 
end up with an edible cake every 

1. East TD. Morris AH. Wallace CJ. 
Clemmer TP. Orme JF Jr. Weaver 
LK. et al. A strategy for development 
of computerized critical care decision 
support systems. Int J Clin Monit 
Comput 199 1-92:8:263-269. 

2. Morris AH, Wallace CJ. Clemmer 
TP, Orme JF Jr. Weaver LK. Dean 




NC, et al. Extracorporeal CO; re- 
moval therapy for adult respiratory 
distress syndrome patients. Respir 
Care 1990;35:224-231. 

Morris: For Dr Wood's information 
and perhaps others', I think the mean 
tidal volume/kilogram of average 
body weight was about 8 mL. 

East: Oh, in our trial? Yes. 

Morris: So, it's conceivable that the 
outcome is related to that. You 
know, permissive hypercapnia is a 
technique. It wasn't called that, but it 
was done at least as early as 1984. 
Some of you may have knowledge of 
references that go further back than 
that but Perret and his colleagues 
from Lausanne reported rather high 
survival for asthma patients 1 treated 
in an institution as patients who were 
quite severely ill requiring intuba- 
tion, and at that time they ignored the 
Pco:- Hickling, in fact, doesn't use a 
lower limit for pH. 2 As Tom East 
displayed, Hickling's lowest pH was 
about 6.8 or 6.9, and he didn't bat an 
eye. We would be uncomfortable at 
this point. 

1. Darioli R. Perret C. Mechanical con- 
trolled hypoventilation in status asth- 
maticus. Am Rev Respir Dis 1984; 

2. Hickling KG, Henderson SJ, Jackson 
R. Low mortality associated with low 
volume, pressure-limited ventilation 
with permissive hypercapnia in severe 
adult respiratory distress syndrome. 
Intensive Care Med 1990;16:372-377. 

East: That's true. 

Wood: Thank you for that infor- 
mation. It doesn't suiprise me that 
the mean tidal volume in your study 
was what it was. My reason for mak- 
ing the point is that anybody who 
thinks they're going to get per- 
missive hypercapnia in ARDS with a 
tidal volume 6-8 mL/kg isn't paying 
attention — those intensivists who use 
12-15 mL/kg don't understand that 
when they turn tidal volume down to 
the levels that you were using, the 
CO: doesn't rise, because the patient 
initiates a faster ventilator rate and 
the alveolar dead space decreases 
with reduced alveolar pressure. 

Morris: Actually, our CO;S were el- 
evated, that's one of the distin- 

guishing features between the recent 
extracorporeal trial and the ECMO 
trial. During the ECMO trial, every- 
body worked hard to keep the Pco; 
normal, but our Pco: average in the 
recent trial was about 56 or some- 
thing like that. So, we did have hy- 
percapnia but no acidemia. 

Wood: Now don't confuse me with 
facts. My mind is made up. 

Hudson: That's not why your sur- 
vival's better. Alan. From the slide 
you showed, you had the patient on a 
kinetic bed. 

Morris: It was Len Hudson who 
picked up on the bed, but he didn't 
notice (or maybe he is saving that 
until tomorrow) that there were ac- 
tually new color TVs for every bed 
in recent clinical trials. 

Hudson: It looked like there was a 
priest in the corner of the room. too. 

Morris: In the 1970s, we had black 
and white TVs. 



Notices of competitions, scholarships, fellowships, examination dates, new educational programs, and the like will be listed here free of 
charge. Items for the Notices section must reach the Journal 60 days before the desired month of publication (January 1 for the March 
issue, February I for the April issue, etc). Include all pertinent information and mail notices to RESPIRATORY CARH Notices Dept, 
1 1030 Abies Lane, Dallas TX 75229-4593. 


The American Respiratory Care Foundation Awards for 1993 

1. $2,000 for the best original paper (study, evaluation, or case report) accepted for publication from November 1992 
through October 1993. This award is not limited to papers based on OPEN Forum presentations. 

2 Four awards of $1,000 each for papers accepted for publication from November 1992 through October 1993 based on 
any Open Forum presentation (not limited to 1992 Open Forum). 

3. Five awards of $500 each for the best papers submitted (not necessarily published) by 1993 OPEN FORUM participants 
who have 'never published' in the Journal. The never-published first author must present the abstract at the Annual 
Meeting and must submit a paper based on the abstract before the 1993 Annual Meeting (received in the Editorial 
Office by November 1, 1993). Co-authors may have previously published in Respiratory Care. 

Three awards of $333 each are to be awarded to the authors of the three best features from Test Your Radiologic Skill, 
Blood Gas Corner, Kittredge's Corner, and PFT Corner accepted for publication from November 1992 through October 
1993. All three (or none) of the features may be chosen from a specific category (eg, all three may be chosen from Blood 
Gas Corner). 

All awards will be made at the 1993 Annual Meeting. Papers are judged automatically. No application is necessary. 

Registration Reimbursement 

Any 1993 Open FORUM presenter (or co-author designee), as in the past, will receive complimentary registration for an 
adequately prepared paper based on his 1993 OPEN FORUM abstract, submitted prior to or at the 1993 Annual Meeting. 

THE NATIONAL BOARD FOR RESPIRATORY CARE— 1993 Examination and Fee Schedule 

CRTT Examination 

Fee Schedule 

Applications Accepted Beginning: 
Application Deadline: 

JULY 17. 1993 

March 1, 1993 

May 1, 1993 

Entry Level CRTT — new applicant: 
Entry Level CRTT — reapplicant: 
RRT Written and Clinical Simulation — 
new applicant: 

$ 90.00 
$ 60.00 


RRT Examination 

Applications Accepted Beginning: 

JUNE 5, 1993 
December 1, 1992 

RRT Written and Clinical Simulation — reapplication: 
Written Registry Only — new applicant: 
Written Registry Only — reapplicant: 
Clinical Simulation Only — new & reapplicant: 

$ 90.00 
$ 60.00 

Application Deadline: 

February 1, 1993 

Entry Level CPFT — new applicant: 


CPFT Examination 

Entry Level CPFT — reapplicant: 
Advanced RPFT — new applicant: 

$ 80.00 


JUNE 5, 1993 

Advanced RPFT — reapplicant: 


Applications Accepted Beginning: 
Application Deadline: 

December 1. 1992 
April I. 1993 

Perinatal/Pediatric — new applicant: 
Perinatal/Pediatric — reappl icant : 


RPFT Examination 



Applications Accepted Beginning: 
Application Deadline: 


July 1, 1993 

September 1, 1993 

CRTT Recredentialing: 
RRT Recredentialing: 
Written Registry Examination 
Clinical Simulation Examination 



$ 60.00 

$ 60.00 

Perinatal/Pediatric Specialty Examination 


CPFT Recredentialing: 
RPFT Recredentialing: 
P/P Spec Credentialing: 


$ 80.00 

Applications Accepted Beginning: 

July 1. 1993 

Membership Renewal: 

Application Deadline: 

November 1. 1993 




8310 Neiman Road • Lenexa, 

Kansas 66214 • (913) 599-4200 



of Events 

Not-for-profit organizations are offered a free advertisement of up to eight lines to appear, on a space-available basis, in Calendar of Events in 
RESPIRATORY CARE Ads for other meetings are priced at $5.50 per line and require an insertion order. Deadline is the 20th of the month two 
months preceding the month you wish the ad to run. Submit copy and insertion orders to Calendar of Events, RESPIRATORY CARE. 1 1030 Abies 
Lane, Dallas TX 75229-4593. 


June 24-26 in Monterey, California. The CSRC presents its 
25th Silver Anniversary Convention, "Expanding the Frontiers 
of Respiratory Care," at the Monterey Marriott Hotel/Monterey 
Conference Center. Keynote speaker David Pierson MD kicks 
off the conference with "Surveying the Frontier: Twenty-Five 
Years of Respiratory Care." Special events include the Sputum 
Bow] finals and a '60s retro dance with awards for the top '60s 
costume. Over 100 companies are exhibiting. Contact John C 
Bigler, Executive Director, CSRC Executive Office, 24307 
Magic Mountain Pkwy, Suite 288, Valencia CA 91355. 
(805) 298-4010, fax (805)298-0235. 

July 9 in Columbus, Ohio. The OSRC presents the 1st Annual 
Adult and Pediatric Critical Care Symposium at the Fawcett 
Center for Tomorrow. The focus of the meeting is new trends in 
sepsis and ARDS. Speakers include James Stoller MD, Herbert 
Weidemann MD, Charles Fisher MD, Jeffrey W Weiland MD, 
and Richard Branson RRT. The conference concludes with case 
studies and a panel discussion. Contact Dennis Giles at (216) 
444-5797, f ax (216) 444-8279. 

July 16-18 in Vail, Colorado. The AARC's Summer Forum, 
featuring education and management programs, is held at the 
Westin Hotel in Vail. For details refer to the special Summer 
Forum Program in the April AARC Times or call (214) 243-2272 
or fax (214) 243-2720. 

July 29 — AARC Videoconference. The AARC. in conjunc- 
tion with VHA Satellite Network, presents the fourth of a six- 
part videoconference series titled "Professor's Rounds in 
Respiratory Care." The fourth presentation, "Monitoring Oxy- 
genation in the Critically 111 Patient," features Leonard D 
Hudson MD and David J Pierson MD. Call (214) 830-0061. 

August 11-13 in Albuquerque, New Mexico. The New 

Mexico Society for Respiratory Care presents its annual con- 
vention at the Albuquerque Convention Center. Highlights 
include lectures by AARC President Dianne Lewis, Thomas 
Petty, Robert Kacmarek, and Louise Nett. Contact Schuyler 
Michael, Pulmonary Rehab. PO Box 26666. Albuquerque NM 

September 22-24 in Frankenmuth, Michigan. The MSRC 
presents its Annual Fall Conference at the Bavarian Inn Motor 
Lodge. The Pulmonary Rehabilitation Membership Section 
opens the conference with a full day of lectures. The following 
two days feature lectures/workshops in pulmonary and sleep 
diagnostics. Concurrent lectures are presented on pediatrics, 
management, gerontology, and adult critical care. Social events 
include a reception, golf outing, sand volleyball, and an outdoor 
steak fry. Exhibitors will be on hand to display the latest in tech- 
nology. Come see Michigan's "Little Bavaria" and enjoy the 
autumn colors of the Midwest. Contact Caroline Kimmel at 

October 15 in Uniondale, New York. The Southeastern 
Chapter of the NYSRC holds its 25th Annual Symposium, "A 
Silver Celebration," at the Marriott Hotel. Speakers and topics 
include Diane Lewis RRT, "The Respiratory Therapist in the 
Year 2001;" Richard Branson RRT. "Essentials of Mechanical 
Ventilator Orders" and "The AARC Consensus on Ventila- 
tors — Third Generation;" Charles Stolar MD, "ECMO and Neo- 
natal Respiratory Distress;" Alan Fein MD, "Tuberculosis and 
Its Impact on Health Care Professionals;" and Judith Tietsort 
RN RRT, "Therapist-Driven Protocols." The symposium fea- 
tures a managers' breakfast, 50 exhibitor booths, the annual 
awards luncheon, and concludes with a "Silver Anniversary" 
dinner dance. Contact Edward Schrager RRT at (7 1 8) 262-6089. 


July 21 in Kansas City, Missouri. An interactive full-day 
workshop. "The Health Professional's Role in Nicotine De- 
pendency Treatment (How to Help Your Patients Stop Smok- 
ing)," includes lectures on "Motivating Smokers to Quit." 
"Evaluating and Treating Nicotine Dependence." and "De- 
veloping a Program to Assist Smokers." Contact Mary Cullen 
CRTT, Pulmonary Rehab Coordinator, Menorah Medical 
Center, 4949 Rockhill Road, Kansas City MO 641 10. (816) 

October 24-29 in Jerusalem, Israel. The XlVth World 
Congress of Asthmalogy convenes in Jerusalem. Topics pre- 
sented include molecular biology in lung disease, cells involved 
in asthma, and the house-dust mite and asthma. Abstracts are 
currently being accepted. Contact Gil-Kenes, Suite 946, 1617 
JFK Blvd. Philadelphia PA 19103. (800) 223-3855. 



News releases about new products and services will be considered tor publication in this section. There is no charge fo 
these listings. Send descriptive release and glossy black and white photographs to RESPIRATORY CARE journal New 
Products and Services Dept. 1 1030 Abies Lane, Dallas TX 75229. 

New Products 
& Services 


The CardiO: integrates simultaneous, 
real-time 12-lead ECG and directly- 
measured Vo : (O: uptake) data in a 
single system, on a single screen. 
According to the manufacturer, if a 
patient has a circulatory problem, the 
physician can pinpoint where it 
lies — the heart, lungs, or blood ves- 
sels — and, in the same time it takes 
to do a standard ECG stress test, 
identify cardiovascular conditions 
that don't show up on ECG alone. 
Medical Graphics Corp, Dept RC, 
350 Oak Grove Pkwy, St Paul MN 


papers emanating from the 1991 
Conference "Breathlessness: The 
Campbell Symposium" held in Ham- 
ilton, Ontario, Canada, are available 
free of charge as a hardback book. 
Topics include Respiratory Muscle 
Action and Respiratory Muscles: 
Structure, Size, and Adaptive Capac- 
ity by Dudley Rochester; The Dia- 
phragm in Emphysema by Peter 
Macklem; Breathlessness in Acute 
Asthma by Sol Permutt; and many 
others. For a copy, write, telephone, 
or fax Mel Freedman, Continuing 
Education Services, Boehringer 
Ingelheim (Canada Ltd), 5180 South 
Service Rd. Burlington. Ontario. 
Canada L7L 5H4; phone (416) 639- 
0333; fax (416) 639-3769. Please 
mention Respiratory Care when 
you make your request. 

POSTER. Invacare Corp offers a 
poster that presents the company's 
Mobilaire™ home respiratory care 
products and recognizes the work of 
home respiratory care practitioners. 
A whimsical takeoff on the movie 
character "Indiana Jones." the poster 
was designed to acknowledge the 
fact that respiratory care practitioners 
face challenges unknown to their in- 
stitution-based colleagues — in order 
to do their job, they, like the char- 
acter, must be resourceful, creative, 
innovative, and able to think on their 
feet. Mobilaire equipment featured 
on the poster is designed to be both 
user-friendly and reliable in the 
home environment, giving therapists 
one less thing to worry about as they 
provide cost-effective care in pa- 
tients' homes. Posters are available 
at no charge to respiratory care 
equipment dealers and home res- 
piratory care practitioners — dealers 
and therapists should call Invacare's 
fullfillment service at (216) 329- 
6456 and request the Adventure 
Poster. Please mention RESPIRATORY 
Care when you call. 

ANALYZER. According to the 
manufacturer, StatPal II is a pH and 
blood gas analysis system that com- 
bines simple operation and true port- 
ability with fast, accurate bedside 
results — 60 seconds after sample 
introduction. StatPal II provides pH, 
P02, and Pco2 measurements in the 
OR, ER, and ICU, or anywhere you 
need quick, reliable answers. The 
unit's reliable sensing electrodes 
reside in its disposable sensor mod- 
ule; calibration and analysis phases 
are easily accomplished, and it is 
CLIA compatible. StatPal II is sup- 
plied with 2 rechargeable battery 
packs that plug into any standard 
outlet. PPG Industries, Biomedical 
Systems Division (Sensors). Dept 
RC. 11077 N Torrey Pines Rd. La 
Jolla CA 92037. (619) 552-9022. 


Vail. Colorado, July 16-18, 1993 


1993 Nashville. Tennessee 

December 11-14 

1994 Las Vegas, Nevada 

December 10-13 

1995 Orlando. Florida 

December 2-5 

1996 San Diego. California 

November 2-5 




in This Issue 

Branson, Richard D 672 

East, Thomas D 690 

Hess, Dean 646 

Higgins, Thomas L 603 

Kacmarek, Robert M 646 

Nelson, Loren D 631 

Phang, P Terry 618 

Pierson, David J 587 

Russell. James A 618 

Yared, Jean-Pierre 603 

in This Issue 

AVL Scientific Corp 581 

Chad Therapeutics 583 

CNS Inc 576 

DHD Medical 574 

Drager Critical Care Systems 570 

HealthScan Products 569 

Invacare Corp 558 

Nellcor 560 

Ohmeda 564, 579 


See Career Opportunities 572A 

Professional Medical Products 580 

Pulsair 566 

Radiometer America Cover 3 

Respironics 573 

Ross Laboratories 563. 564 

Sherwood Medical Cover 4 

Siemens Medical System Cover 2 

Transtracheal Systems 575 



For faster 

service, FAX 

your reader 

service card to 


81 AARC 

Membership Information 

82 Respiratory Care 
Subscription Information 

143 AVL Scientific Corp 
Blood Gas Analyzer 

109 Chad Therapeutics 

Portable Oxygen 

117 CNSInc 

Sleep Testing 
131 DHD Medical 

ACE (Aerosol Cloud 


101 Drager Critical Care 
Minimum Pressure 

136 HealthScan Products 
Assess Peak Flow Meter 

123 Invacare Corp 
Passport Aerosol 

113 Nellcor 

Pulse Oximeters 

103 Ohmeda 

Pulse Oximetry Products 

144 Ohmeda 
Capnography Videotape 

134 Professional Medical 
(3Mist2 Nebulizer 

102 Pulsair 

Oxygen Conservation 

124 Radiometer America 
Blood Gas, Hemoglobin, 
& Electrolyte Analyzers 

129 Respironics Inc 

BiPAP S/T-D System 

125 Ross Laboratories 

155 Sherwood Medical 
Incentive Deep 
Breathing Exerciser 

114 Siemens Medical 

Servo Ventilator 300 

104 Transtracheal Systems 
Transtracheal Oxygen 


Information Requests or 
Change of Address 

Please complete the card below 

AARC Membership No. 

Old Address 




New Address 



Check the boxes below 
for information from the 

□ Change of address 

□ AARC Membership 

□ AARC Catalog 

□ AARC Position 


June Reader Service 

Expires September 30, 1993 



Please Circle No More Than 15 Items 


























































































































June Reader Service 

Expires September 30, 1993 

Street _ 


Please Circle No More Than 15 Items 

























































































































I. Type of Instn/Practice 

1 J Hosp 2 500 or more bed 

2 J Hosp 300 to 499 beds 

3 J Hosp 200 to 299 beds 

4 J Hosp 100 to 199 beds 

5 J Hosp <l00ot less bed 

6 J Skilled Nursing Facility 

7 _l Home Care Practice 

8 J School 

It. Department 
A.J Respiratory Therapy 
B.J Cardiopulmonary 
C. J Anesthesia Service 
D.J Emergency Dept 

1 J Clinical Practice 

2 J Pennatal Pediatncs 

3 J Cntical Care 

4 J Clinical Research 

5 J Pulmonary Function Lat 

6 J Home Care/Rehab 

8 J Management 
IV. Position 
A.J Dept Head 
B.J Chief Therapist 
C.J Supervisor 
D.J Staff Technician 
E J Staff Therapist 
F.J Educator 
G J Medical Director 
H J Anesthesiologist 
I J Pulmonologisl 
J J Other MD 

t. Type of Instn/Practice 

I . J Hosp > 500 or more bed 

2 J Hosp 300 to 499 beds 

3 J Hosp 200 to 299 beds 

4 J Hosp 100 to 199 beds 

5 J Hosp <100 or less beds 

6 J Skilled Nursing Facility 

7 J Home Care Practice 

8 J School 

II. Department 

A.J Respiratory Therapy 
B J Cardiopulmonary 
C J Anesthesia Service 
D.J Emergency Dept 

III. Specialty 

I.J Clinical Practice 

2 J Pennatal Pediatncs 

3 J Cntical Care 

4 J Clinical Research 

5 J Pulmonary Function Lat 

6 J Home Care/Rehab 

7 J Education 

8 J Management 

IV. Position 

A J Dept Head 
B J Chief Therapist 
C.J Supervisor 
D.J Staff Technician 
E J Staff Therapist 
F.J Educator 
G.J Medical Director 
H J Anesthesiologist 
I J Pulmonologisl 
J J Other MD 

I Nur* 

rot t 





P.O. BOX 29686 

DALLAS, TX 75229-9691 






ll...l,l,l,,.l.l..l,ll,l.,l,l,„ll..l,l 1 1 1 1 ... I 




P.O. BOX 1856 
RIVERTON, NJ 08077-9456 





UNITED STATES,l.l„ll„.l..ll 




P.O. BOX 1856 
RIVERTON, NJ 08077-9456 






• l»l.ll...l,„ll...ll.l,.,l„l.l.l,.ll„.l..ll 


pH » pCO, pO, 


Expand on traditional blood gas analysis. During critical 
care treatment, The Deep Picture provides the basis for 
therapy aimed at improving the oxygen supply capacity of 
arterial blood. Add electrolyte analysis and expand The 
Deep Picture for the exact set of measurements you need. 

For important information 

about CLIA '88 Regulations 

call Radiometer America, Inc 


ext. 223. 

Formore information aboutTheDeepPictureandRadiometer 



8 1 1 Sharon Drive, Westlake, OH 44 1 45 

Circle 124 on reader service card 


The Leader in Blood Gas Systems 


Volumetric Incentive Deep-Breathing Exerciser 

The accuracy of Voldyne, in a new size, matched to geriatric 
patients and patients with smaller lung capacities. 



A smaller, lighter flow cup reduces the work of breathing, thus 
improving patient performance and progress. 

■ Every unit is individually tested and calibrated for performance, 
reliability and superior accuracy of inhaled lung volume. 

■ Volume incentive spirometry improves assessment of patient 

progress by eliminating the guesswork associated with spirometers 
that only measure flow. 

■ Graduations printed on both sides of the unit allow the therapist to 
conveniently observe volumes while instructing and encouraging 
the patient. 

For further information, contact uoar Sherwood OR. /Critical Care 
Representative or call n at* 

I-oUU-325-7472 (outside Missouri) 

1-800-392-7318 (in Missouri) 

©1991 Sherwood Medical Company 

A Sherwood 


Circle 155 on reader service card